# AUTO-GENERATED by scripts/build_concepts_yaml.py — DO NOT EDIT BY HAND. # Single source of truth for canonical concepts, derived from # CONCEPT: markers in agent_utilities/ (*.py, *.rs). generated_by: scripts/build_concepts_yaml.py total_concepts: 319 total_pillars: 20 pillars: - AHE-3 - CE-038 - CTX-1 - ECO-4 - EE-033 - EE-034 - EE-036 - EE-037 - EE-039 - EG-009 - KG-1 - KG-2 - LGC-1 - ML-011 - ORCH-1 - ORCH-2 - ORCH-5 - OS-5 - SAFE-1 - UTIL-1 concepts: - id: AHE-3.x name: Telemetry-Driven Optimization pillar: AHE-3 status: live code_paths: - agent_utilities/graph/_router_impl.py - agent_utilities/models/schema_definition.py doc: Telemetry-Driven Optimization - id: AHE-3.0 name: Agentic Harness Engineering / Evolution pillar: AHE-3 status: live code_paths: - agent_utilities/harness/agentic_evolution_engine.py - agent_utilities/harness/component_registry.py - agent_utilities/harness/constraint_engine.py - agent_utilities/harness/continuous_evaluation_engine.py - agent_utilities/harness/evidence_corpus.py - agent_utilities/harness/evolve_agent.py - agent_utilities/harness/manifest.py - agent_utilities/harness/replay_buffer.py - agent_utilities/harness/trace_backend.py - agent_utilities/harness/verifier.py - agent_utilities/knowledge_graph/backends/owl/owlready2_backend.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/kb/x_workflows.py - agent_utilities/knowledge_graph/orchestration/engine_ahe.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/schema_definition.py - agent_utilities/models/schema_packs/evolution.py - agent_utilities/rlm/config.py - agent_utilities/sdd/__init__.py doc: Agentic Harness Engineering / Evolution - id: AHE-3.1 name: / AHE-3.40 — generalized from the original system-prompt-only pillar: AHE-3 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/capabilities/adversarial_verifier.py - agent_utilities/core/config.py - agent_utilities/graph/__init__.py - agent_utilities/graph/horizon_curriculum.py - agent_utilities/graph/parallel_engine.py - agent_utilities/graph/reward_decomposition.py - agent_utilities/graph/training_signals.py - agent_utilities/graph/verification.py - agent_utilities/harness/__init__.py - agent_utilities/harness/citation_tracker.py - agent_utilities/harness/continuous_evaluation_engine.py - agent_utilities/harness/eval_corpus.py - agent_utilities/harness/evaluation_engine.py - agent_utilities/harness/evolve_agent.py - agent_utilities/harness/imodel_evolver.py - agent_utilities/harness/reasoning_effort.py - agent_utilities/harness/red_team.py - agent_utilities/harness/reliability_corpus.py - agent_utilities/harness/reliability_scorers.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/memory/optimization_engine.py - agent_utilities/models/imodel.py - agent_utilities/models/knowledge_graph.py - agent_utilities/prompting/builder.py - agent_utilities/prompting/dspy_compiler.py - agent_utilities/tools/eval_harness.py doc: / AHE-3.40 — generalized from the original system-prompt-only - id: AHE-3.2 name: Optional convergence monitor for multi-loop tasks pillar: AHE-3 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/core/cognitive_scheduler.py - agent_utilities/graph/hierarchical_planner.py - agent_utilities/graph/population_drift.py - agent_utilities/graph/workspace_attention.py - agent_utilities/harness/agentic_evolution_engine.py - agent_utilities/harness/engineering.py - agent_utilities/harness/self_guided_play.py - agent_utilities/harness/variant_pool.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/models/imodel.py - agent_utilities/models/knowledge_graph.py - agent_utilities/observability/config_versioning.py - agent_utilities/patterns/__init__.py - agent_utilities/patterns/exploration.py - agent_utilities/patterns/first_run_tests.py - agent_utilities/tools/self_improvement_tools.py - agent_utilities/workflows/distillation_hook.py doc: Optional convergence monitor for multi-loop tasks - id: AHE-3.3 name: Check for matching TeamConfig before LLM planning pillar: AHE-3 status: live code_paths: - agent_utilities/capabilities/teams.py - agent_utilities/core/registry/kg_adapter.py - agent_utilities/graph/_router_impl.py - agent_utilities/graph/routing/strategies/team_reuse.py - agent_utilities/graph/subagent_patterns.py - agent_utilities/graph/verification.py - agent_utilities/harness/continuous_evaluation_engine.py - agent_utilities/harness/imodel_evolver.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/retrieval/memory_retriever.py - agent_utilities/models/imodel.py - agent_utilities/models/knowledge_graph.py - agent_utilities/security/execution_stability_engine.py - agent_utilities/tools/team_tools.py - agent_utilities/workflows/distillation_hook.py doc: Check for matching TeamConfig before LLM planning - id: AHE-3.4 name: Detected mathematical/quantitative topology. Escalate to reasoning model pillar: AHE-3 status: live code_paths: - agent_utilities/core/agentspec_catalog.py - agent_utilities/graph/__init__.py - agent_utilities/graph/_router_impl.py - agent_utilities/graph/hierarchical_planner.py - agent_utilities/graph/horizon_curriculum.py - agent_utilities/graph/reward_decomposition.py - agent_utilities/graph/team_evolution.py - agent_utilities/graph/verification.py - agent_utilities/graph/workspace_attention.py - agent_utilities/harness/__init__.py - agent_utilities/harness/background_spawner.py - agent_utilities/harness/continuous_evaluation_engine.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/memory/optimization_engine.py - agent_utilities/knowledge_graph/memory/strategies/heavy_thinking_cache.py - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/schema_packs/finance.py - agent_utilities/patterns/manual_testing.py - agent_utilities/security/execution_stability_engine.py doc: Detected mathematical/quantitative topology. Escalate to reasoning model - id: AHE-3.9 name: Distills updated tool description back into Python function docstring pillar: AHE-3 status: live code_paths: - agent_utilities/knowledge_graph/distillation/physical_distiller.py - agent_utilities/knowledge_graph/distillation/skill_graph_distiller.py - agent_utilities/knowledge_graph/distillation/skill_graph_importer.py - agent_utilities/mcp/tools/write_ingest_tools.py doc: Distills updated tool description back into Python function docstring - id: AHE-3.11 name: GitOps Git Commit Automation pillar: AHE-3 status: live code_paths: - agent_utilities/knowledge_graph/distillation/physical_distiller.py doc: GitOps Git Commit Automation - id: AHE-3.12 name: LongMemEval-S validation harness pillar: AHE-3 status: live code_paths: - agent_utilities/server/app.py - agent_utilities/server/routers/benchmark.py doc: LongMemEval-S validation harness - id: AHE-3.13 name: pre-emit quality gate on the live execution path pillar: AHE-3 status: live code_paths: - agent_utilities/core/execution/engine.py - agent_utilities/harness/__init__.py - agent_utilities/harness/provenance_gate.py - agent_utilities/harness/quality_gates.py - agent_utilities/harness/self_guided_play.py doc: pre-emit quality gate on the live execution path - id: AHE-3.14 name: consider promoting the team pillar: AHE-3 status: live code_paths: - agent_utilities/knowledge_graph/research/auto_merge.py - agent_utilities/knowledge_graph/research/loop_controller.py doc: consider promoting the team - id: AHE-3.15 name: this is the read-back half of agent-step policy pillar: AHE-3 status: live code_paths: - agent_utilities/graph/agent_step_po.py - agent_utilities/graph/reward_decomposition.py - agent_utilities/graph/routing/strategies/policy.py - agent_utilities/orchestration/agent_runner.py doc: this is the read-back half of agent-step policy - id: AHE-3.16 name: width of the *diverse* best-of-k fan-out — harder pillar: AHE-3 status: live code_paths: - agent_utilities/graph/test_time_diversity.py - agent_utilities/harness/reasoning_effort.py doc: width of the *diverse* best-of-k fan-out — harder - id: AHE-3.17 name: Preference-Corpus Reliability pillar: AHE-3 status: live code_paths: - agent_utilities/harness/preference_pairs.py - agent_utilities/knowledge_graph/adaptation/feedback.py doc: Preference-Corpus Reliability - id: AHE-3.18 name: Failure-Driven Evolution pillar: AHE-3 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/graph/parallel_engine.py - agent_utilities/harness/trace_backend.py - agent_utilities/knowledge_graph/adaptation/failure_analyzer.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/research/change_publisher.py - agent_utilities/knowledge_graph/research/loop_controller.py - agent_utilities/mcp/tools/analysis_tools.py doc: Failure-Driven Evolution - id: AHE-3.19 name: Performance Anomaly Consumer pillar: AHE-3 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/adaptation/anomaly_consumer.py - agent_utilities/knowledge_graph/adaptation/failure_analyzer.py - agent_utilities/knowledge_graph/core/engine_tasks.py doc: Performance Anomaly Consumer - id: AHE-3.20 name: Promotion Governance Validator pillar: AHE-3 status: live code_paths: - agent_utilities/knowledge_graph/adaptation/failure_analyzer.py - agent_utilities/knowledge_graph/research/auto_merge.py - agent_utilities/knowledge_graph/research/promotion_governance.py doc: Promotion Governance Validator - id: AHE-3.21 name: Evolution-to-branch bridge publishing promoted proposals as regression-gated reviewable local git branches via change synthesis, RLM-sandbox validation and a governed ChangePublisher seam pillar: AHE-3 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/research/auto_merge.py - agent_utilities/knowledge_graph/research/change_publisher.py - agent_utilities/knowledge_graph/research/change_synthesis.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/orchestration/fleet_reconciler.py doc: Evolution-to-branch bridge publishing promoted proposals as regression-gated reviewable local git branches via change synthesis, RLM-sandbox validation and a governed ChangePublisher seam - id: AHE-3.22 name: autonomous single-file code-synthesis stage that generates a real diff for an attributed promoted proposal so the deployed evolution loop emits code instead of only a prose plan pillar: AHE-3 status: live code_paths: - agent_utilities/harness/swebench_corpus.py - agent_utilities/harness/swebench_harness.py - agent_utilities/knowledge_graph/research/change_publisher.py - agent_utilities/knowledge_graph/research/change_synthesis.py - agent_utilities/knowledge_graph/research/code_synthesis.py - agent_utilities/server/routers/swebench.py doc: autonomous single-file code-synthesis stage that generates a real diff for an attributed promoted proposal so the deployed evolution loop emits code instead of only a prose plan - id: AHE-3.23 name: verified apply-verify-rollback from the ManifestVerifier confirm-or-revert recommendation over measured before-and-after capability scores pillar: AHE-3 status: live code_paths: - agent_utilities/harness/swebench_remediation.py - agent_utilities/knowledge_graph/research/capability_ratchet.py - agent_utilities/knowledge_graph/research/change_publisher.py doc: verified apply-verify-rollback from the ManifestVerifier confirm-or-revert recommendation over measured before-and-after capability scores - id: AHE-3.24 name: capability-benchmark regression ratchet re-measuring a published worktree against a persisted baseline and abandoning the branch on any tracked-capability regression pillar: AHE-3 status: live code_paths: - agent_utilities/knowledge_graph/research/capability_ratchet.py doc: capability-benchmark regression ratchet re-measuring a published worktree against a persisted baseline and abandoning the branch on any tracked-capability regression - id: AHE-3.25 name: Plain-English regression assertions with verified-remediation auto-lock pillar: AHE-3 status: live code_paths: - agent_utilities/harness/continuous_evaluation_engine.py - agent_utilities/harness/eval_corpus.py - agent_utilities/knowledge_graph/adaptation/failure_analyzer.py doc: Plain-English regression assertions with verified-remediation auto-lock - id: AHE-3.26 name: recursive-improvement instrumentation aggregating per-cycle proposal merge and capability-delta metrics so the self-evolution loop can measure its own rate and per-mechanism contribution pillar: AHE-3 status: live code_paths: - agent_utilities/knowledge_graph/research/improvement_ledger.py - agent_utilities/knowledge_graph/research/loop_controller.py doc: recursive-improvement instrumentation aggregating per-cycle proposal merge and capability-delta metrics so the self-evolution loop can measure its own rate and per-mechanism contribution - id: AHE-3.27 name: per-task adaptation-speed instrumentation pillar: AHE-3 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/adaptation_speed.py - agent_utilities/harness/hote_tri_evolution.py doc: per-task adaptation-speed instrumentation - id: AHE-3.28 name: the pillar: AHE-3 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/sai_task.py doc: the - id: AHE-3.29 name: the controller that operationalizes Superhuman Adaptable pillar: AHE-3 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/research/sai_factory.py - agent_utilities/mcp/tools/analysis_tools.py doc: the controller that operationalizes Superhuman Adaptable - id: AHE-3.30 name: ────────────────── pillar: AHE-3 status: live code_paths: - agent_utilities/graph/training_signals.py - agent_utilities/knowledge_graph/search_synthesis/__init__.py doc: ────────────────── - id: AHE-3.31 name: a recursive-distillation loop that fine-tunes a new prior from the search-distilled corpus, gates the candidate model behind a monotone capability ratchet, and promotes it only on a non-regressing improvement so test-time search converts into a stronger next prior pillar: AHE-3 status: live code_paths: - agent_utilities/harness/recursive_distill.py - agent_utilities/mcp/tools/analysis_tools.py doc: a recursive-distillation loop that fine-tunes a new prior from the search-distilled corpus, gates the candidate model behind a monotone capability ratchet, and promotes it only on a non-regressing improvement so test-time search converts into a stronger next prior - id: AHE-3.32 name: Long-Context RLM Benchmark Harness and Paper-Comparison Scoreboard pillar: AHE-3 status: live code_paths: - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/rlm/benchmarks/__init__.py - agent_utilities/rlm/benchmarks/base.py - agent_utilities/rlm/benchmarks/baselines.py - agent_utilities/rlm/benchmarks/cost.py - agent_utilities/rlm/benchmarks/runner.py - agent_utilities/rlm/benchmarks/scoreboard.py - agent_utilities/rlm/benchmarks/tasks/__init__.py - agent_utilities/rlm/benchmarks/tasks/_datasets.py - agent_utilities/rlm/benchmarks/tasks/browsecomp_plus.py - agent_utilities/rlm/benchmarks/tasks/longbench_codeqa.py - agent_utilities/rlm/benchmarks/tasks/oolong.py - agent_utilities/rlm/benchmarks/tasks/oolong_pairs.py - agent_utilities/rlm/benchmarks/tasks/s_niah.py - agent_utilities/rlm/predict_rlm.py - agent_utilities/rlm/repl.py - agent_utilities/rlm/runner.py doc: Long-Context RLM Benchmark Harness and Paper-Comparison Scoreboard - id: AHE-3.33 name: one online exploit/explore bandit router per agent pillar: AHE-3 status: live code_paths: - agent_utilities/harness/decentralized_memory.py - agent_utilities/harness/explore_exploit_router.py doc: one online exploit/explore bandit router per agent - id: AHE-3.34 name: operationalizes the research-craft habit of *forecasting every pillar: AHE-3 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/continuous_evaluation_engine.py - agent_utilities/harness/forecasting.py doc: operationalizes the research-craft habit of *forecasting every - id: AHE-3.35 name: AHE-3.35 pillar: AHE-3 status: live code_paths: - agent_utilities/harness/baseline_overfit_gate.py doc: AHE-3.35 - id: AHE-3.36 name: disconfirming-evidence research log, fed each round pillar: AHE-3 status: live code_paths: - agent_utilities/harness/continuous_evaluation_engine.py - agent_utilities/harness/research_log.py doc: disconfirming-evidence research log, fed each round - id: AHE-3.37 name: when a task is in scope, run a short self-guided pillar: AHE-3 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/agentic_evolution_engine.py - agent_utilities/harness/self_guided_play.py doc: when a task is in scope, run a short self-guided - id: AHE-3.38 name: Evolvable sampling profiles via parametric variant mutation scored by the capability-reward EMA and tournament-promoted per task-class into the live model registry pillar: AHE-3 status: live code_paths: - agent_utilities/agent/sampling_profile.py - agent_utilities/harness/variant_pool.py - agent_utilities/models/model_registry.py doc: Evolvable sampling profiles via parametric variant mutation scored by the capability-reward EMA and tournament-promoted per task-class into the live model registry - id: AHE-3.39 name: the real optimization metric pillar: AHE-3 status: live code_paths: - agent_utilities/harness/dspy_optimization.py doc: the real optimization metric - id: AHE-3.40 name: Optimizable-target registry + generalized DSPy driver pillar: AHE-3 status: live code_paths: - agent_utilities/harness/dspy_optimization.py - agent_utilities/harness/evolve_agent.py - agent_utilities/knowledge_graph/dspy_kg_bridge.py - agent_utilities/mcp/tools/analysis_tools.py doc: Optimizable-target registry + generalized DSPy driver - id: AHE-3.41 name: MCP tool descriptions become a DSPy-optimizable target pillar: AHE-3 status: live code_paths: - agent_utilities/harness/dspy_optimization.py doc: MCP tool descriptions become a DSPy-optimizable target - id: AHE-3.42 name: agent skill SOP/description as a DSPy-optimizable target; the SOP pillar: AHE-3 status: live code_paths: - agent_utilities/harness/dspy_optimization.py doc: agent skill SOP/description as a DSPy-optimizable target; the SOP - id: AHE-3.43 name: Few-shot demo-set refinement pillar: AHE-3 status: live code_paths: - agent_utilities/harness/dspy_optimization.py doc: Few-shot demo-set refinement - id: AHE-3.44 name: DSPy optimization of the KG extraction prompt against a pillar: AHE-3 status: live code_paths: - agent_utilities/knowledge_graph/extraction/extraction_optimizer.py doc: DSPy optimization of the KG extraction prompt against a - id: AHE-3.45 name: Concept-matching and routing-policy optimization pillar: AHE-3 status: live code_paths: - agent_utilities/harness/policy_optimization.py doc: Concept-matching and routing-policy optimization - id: AHE-3.46 name: a candidate replaces the incumbent only when it pillar: AHE-3 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/harness/dspy_optimization.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/mcp/tools/analysis_tools.py doc: a candidate replaces the incumbent only when it - id: AHE-3.47 name: empirical parity evidence for the assimilation program pillar: AHE-3 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/assimilation_benchmark.py - agent_utilities/mcp/tools/analysis_tools.py doc: empirical parity evidence for the assimilation program - id: AHE-3.48 name: empirical evidence that the latent-native enhancements move pillar: AHE-3 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/latent_efficiency_benchmark.py - agent_utilities/mcp/tools/analysis_tools.py doc: empirical evidence that the latent-native enhancements move - id: AHE-3.49 name: ────────────────────────── pillar: AHE-3 status: live code_paths: - agent_utilities/graph/training_signals.py doc: ────────────────────────── - id: AHE-3.50 name: co-evolve the research pillar: AHE-3 status: live code_paths: - agent_utilities/harness/hote_tri_evolution.py - agent_utilities/knowledge_graph/research/loop_controller.py doc: co-evolve the research - id: AHE-3.51 name: reasoning materialises the variant↔base and edit↔pathology pillar: AHE-3 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py doc: reasoning materialises the variant↔base and edit↔pathology - id: AHE-3.52 name: run the AEGIS loop over a provided edit sequence pillar: AHE-3 status: live code_paths: - agent_utilities/harness/aegis_loop.py - agent_utilities/mcp/tools/analysis_tools.py doc: run the AEGIS loop over a provided edit sequence - id: AHE-3.53 name: the parity-and-surpass scoreboard vs HarnessX pillar: AHE-3 status: live code_paths: - agent_utilities/harness/harness_foundry_benchmark.py - agent_utilities/harness/harness_gate.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/analysis_tools.py doc: the parity-and-surpass scoreboard vs HarnessX - id: AHE-3.55 name: AHE-3.56 pillar: AHE-3 status: live code_paths: - agent_utilities/harness/co_evolution.py doc: AHE-3.56 - id: AHE-3.56 name: /KG-2.108 — held-out certification + ARA-Seal of a pillar: AHE-3 status: live code_paths: - agent_utilities/harness/harness_grounding.py - agent_utilities/mcp/tools/analysis_tools.py doc: /KG-2.108 — held-out certification + ARA-Seal of a - id: AHE-3.57 name: per-dimension ship-outcome ledger pillar: AHE-3 status: live code_paths: - agent_utilities/harness/aegis_loop.py doc: per-dimension ship-outcome ledger - id: AHE-3.58 name: an edit claims a fix via a pillar: AHE-3 status: live code_paths: - agent_utilities/harness/harness_foundry_benchmark.py - agent_utilities/harness/manifest.py - agent_utilities/harness/verifier.py doc: an edit claims a fix via a - id: AHE-3.59 name: ─────────────────────────────────── pillar: AHE-3 status: live code_paths: - agent_utilities/harness/aegis_loop.py - agent_utilities/harness/harness_foundry_benchmark.py doc: ─────────────────────────────────── - id: AHE-3.60 name: '** manifest-verify →' pillar: AHE-3 status: live code_paths: - agent_utilities/harness/aegis_loop.py - agent_utilities/harness/manifest.py doc: '** manifest-verify →' - id: CE-038 name: periodic code-health pillar: CE-038 status: live code_paths: - agent_utilities/knowledge_graph/adaptation/code_health.py doc: periodic code-health - id: CTX-1.0 name: Nested Subfolder Instructions pillar: CTX-1 status: live code_paths: - agent_utilities/nested_context.py doc: Nested Subfolder Instructions - id: ECO-4.0 name: Register PlannerGraphSkill when graph_bundle is available pillar: ECO-4 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/capabilities/manager.py - agent_utilities/core/config.py - agent_utilities/core/paths.py - agent_utilities/core/registry/kg_adapter.py - agent_utilities/core/registry/package_adapter.py - agent_utilities/core/registry/plugin_adapter.py - agent_utilities/domains/finance/quant_mcp_tools.py - agent_utilities/ecosystem/__init__.py - agent_utilities/ecosystem/hook_installer.py - agent_utilities/gateway/graph_api.py - agent_utilities/graph/adaptive_provisioning.py - agent_utilities/graph/builder.py - agent_utilities/graph/hierarchical_planner.py - agent_utilities/graph/parallel_engine.py - agent_utilities/graph/tool_resolver.py - agent_utilities/knowledge_graph/adaptation/skill_evolver.py - agent_utilities/knowledge_graph/core/ecosystem_topology.py - agent_utilities/knowledge_graph/core/engine_ingestion.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/kb/x_ingestion.py - agent_utilities/knowledge_graph/kb/x_workflows.py - agent_utilities/knowledge_graph/memory/cli.py - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/mcp/__init__.py - agent_utilities/mcp/action_dispatch.py - agent_utilities/mcp/concurrency.py - agent_utilities/mcp/context_helpers.py - agent_utilities/mcp/delegated_auth.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/multiplexer.py - agent_utilities/mcp/server_factory.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/mcp_utilities.py - agent_utilities/messaging/__init__.py - agent_utilities/messaging/backends/__init__.py - agent_utilities/messaging/backends/discord.py - agent_utilities/messaging/backends/googlechat.py - agent_utilities/messaging/backends/googlemeet.py - agent_utilities/messaging/backends/imessage.py - agent_utilities/messaging/backends/irc.py - agent_utilities/messaging/backends/line.py - agent_utilities/messaging/backends/matrix.py - agent_utilities/messaging/backends/mattermost.py - agent_utilities/messaging/backends/nextcloud.py - agent_utilities/messaging/backends/signal.py - agent_utilities/messaging/backends/slack.py - agent_utilities/messaging/backends/synology.py - agent_utilities/messaging/backends/teams.py - agent_utilities/messaging/backends/telegram.py - agent_utilities/messaging/backends/twitch.py - agent_utilities/messaging/backends/voicecall.py - agent_utilities/messaging/backends/whatsapp.py - agent_utilities/messaging/base.py - agent_utilities/messaging/capabilities.py - agent_utilities/messaging/kg_ingest.py - agent_utilities/messaging/models.py - agent_utilities/messaging/registry.py - agent_utilities/messaging/router.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/schema_packs/finance.py - agent_utilities/orchestration/durable_execution.py - agent_utilities/prompting/provider_adapter.py - agent_utilities/protocols/a2a.py - agent_utilities/protocols/a2a_config.py - agent_utilities/protocols/a2a_graph_skill.py - agent_utilities/protocols/acp_adapter.py - agent_utilities/protocols/acp_providers.py - agent_utilities/security/permissions_kernel.py - agent_utilities/server/app.py - agent_utilities/server/routers/__init__.py - agent_utilities/server/routers/agent_ui.py - agent_utilities/server/routers/core.py - agent_utilities/tools/agent_tools.py - agent_utilities/tools/browser/browser_control.py - agent_utilities/tools/browser/browser_manager.py - agent_utilities/tools/developer_tools.py - agent_utilities/tools/dynamic_tool_orchestrator.py - agent_utilities/tools/git_tools.py - agent_utilities/tools/jupyter_adapter.py - agent_utilities/tools/onboarding_tools.py - agent_utilities/tools/sandbox_executor.py - agent_utilities/tools/tool_filtering.py - agent_utilities/tools/tool_registry.py - agent_utilities/tools/workspace_tools.py doc: Register PlannerGraphSkill when graph_bundle is available - id: ECO-4.1 name: Live MCP server connection for tool metadata caching pillar: ECO-4 status: live code_paths: - agent_utilities/knowledge_graph/core/engine_mcp_discovery.py - agent_utilities/mcp/harness_server.py - agent_utilities/mcp/middlewares.py doc: Live MCP server connection for tool metadata caching - id: ECO-4.3 name: Company Infrastructure Orchestration pillar: ECO-4 status: live code_paths: - agent_utilities/models/company.py - agent_utilities/models/knowledge_graph.py doc: Company Infrastructure Orchestration - id: ECO-4.04 name: Infrastructure Blueprint Library pillar: ECO-4 status: live code_paths: - agent_utilities/models/company.py - agent_utilities/models/knowledge_graph.py doc: Infrastructure Blueprint Library - id: ECO-4.05 name: Pluggable Event Queue Backend pillar: ECO-4 status: live code_paths: - agent_utilities/knowledge_graph/core/event_backend.py - agent_utilities/knowledge_graph/core/kafka_queue_backend.py - agent_utilities/knowledge_graph/core/queue_backend.py doc: Pluggable Event Queue Backend - id: ECO-4.10 name: Team-Specific Startup Context pillar: ECO-4 status: live code_paths: - agent_utilities/knowledge_graph/memory/memory_engine.py doc: Team-Specific Startup Context - id: ECO-4.11 name: Deterministic Lint Enforcement Hook pillar: ECO-4 status: live code_paths: - agent_utilities/ecosystem/lint_enforcement_hook.py doc: Deterministic Lint Enforcement Hook - id: ECO-4.12 name: Plugin Bundle Distribution System pillar: ECO-4 status: live code_paths: - agent_utilities/ecosystem/plugin_bundle.py doc: Plugin Bundle Distribution System - id: ECO-4.13 name: Permission Policy Engine pillar: ECO-4 status: live code_paths: - agent_utilities/ecosystem/permission_policy.py doc: Permission Policy Engine - id: ECO-4.17 name: Self-Improving AGENTS.md Reflector pillar: ECO-4 status: live code_paths: - agent_utilities/ecosystem/agents_md_reflector.py doc: Self-Improving AGENTS.md Reflector - id: ECO-4.21 name: Configuration Staleness Auditor pillar: ECO-4 status: live code_paths: - agent_utilities/ecosystem/config_staleness_auditor.py doc: Configuration Staleness Auditor - id: ECO-4.22 name: Governance Workflow Pipeline pillar: ECO-4 status: live code_paths: - agent_utilities/ecosystem/agent_manager_dashboard.py - agent_utilities/ecosystem/governance_workflow.py doc: Governance Workflow Pipeline - id: ECO-4.23 name: Codebase Map Generator pillar: ECO-4 status: live code_paths: - agent_utilities/tools/codebase_map_tools.py doc: Codebase Map Generator - id: ECO-4.24 name: Langfuse exporter for spans, token usage and traces pillar: ECO-4 status: live code_paths: - agent_utilities/observability/langfuse_exporter.py - agent_utilities/orchestration/engine.py doc: Langfuse exporter for spans, token usage and traces - id: ECO-4.25 name: Document-Source Connector Framework pillar: ECO-4 status: live code_paths: - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/protocols/source_connectors/__init__.py - agent_utilities/protocols/source_connectors/base.py - agent_utilities/protocols/source_connectors/connectors/__init__.py - agent_utilities/protocols/source_connectors/connectors/database.py - agent_utilities/protocols/source_connectors/connectors/filesystem.py - agent_utilities/protocols/source_connectors/connectors/rest.py - agent_utilities/protocols/source_connectors/connectors/web.py doc: Document-Source Connector Framework - id: ECO-4.26 name: incremental poll keyed on a crawl-time watermark + seen ids pillar: ECO-4 status: live code_paths: - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/protocols/source_connectors/base.py - agent_utilities/protocols/source_connectors/checkpoint.py - agent_utilities/protocols/source_connectors/connectors/database.py - agent_utilities/protocols/source_connectors/connectors/filesystem.py - agent_utilities/protocols/source_connectors/connectors/mcp_package.py - agent_utilities/protocols/source_connectors/connectors/rest.py - agent_utilities/protocols/source_connectors/connectors/web.py doc: incremental poll keyed on a crawl-time watermark + seen ids - id: ECO-4.27 name: the discovery half of the self-registering pattern. Called pillar: ECO-4 status: live code_paths: - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/mcp/kg_server.py - agent_utilities/protocols/source_connectors/connectors/__init__.py - agent_utilities/protocols/source_connectors/registry.py doc: the discovery half of the self-registering pattern. Called - id: ECO-4.28 name: External Permission Sync pillar: ECO-4 status: live code_paths: - agent_utilities/protocols/source_connectors/base.py - agent_utilities/protocols/source_connectors/connectors/filesystem.py - agent_utilities/protocols/source_connectors/permission_sync.py doc: External Permission Sync - id: ECO-4.29 name: declarative catalog driving the MCP fleet adapter pillar: ECO-4 status: live code_paths: - agent_utilities/protocols/source_connectors/connectors/mcp_package.py - agent_utilities/protocols/source_connectors/connectors/package_manifest.py doc: declarative catalog driving the MCP fleet adapter - id: ECO-4.30 name: /4.31 — media generation + transcription tools. Default OFF pillar: ECO-4 status: live code_paths: - agent_utilities/ecosystem/media/__init__.py - agent_utilities/ecosystem/media/gateway.py - agent_utilities/ecosystem/media/transcription.py - agent_utilities/tools/media_tools.py - agent_utilities/tools/tool_registry.py doc: /4.31 — media generation + transcription tools. Default OFF - id: ECO-4.31 name: Media Transcription Bridge pillar: ECO-4 status: live code_paths: - agent_utilities/ecosystem/media/transcription.py - agent_utilities/tools/media_tools.py doc: Media Transcription Bridge - id: ECO-4.32 name: apply any query-analysis source-type restriction to pillar: ECO-4 status: live code_paths: - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/knowledge_graph/retrieval/query_analysis.py doc: apply any query-analysis source-type restriction to - id: ECO-4.33 name: native database traversal tools pillar: ECO-4 status: live code_paths: - agent_utilities/tools/db_tools.py - agent_utilities/tools/tool_registry.py doc: native database traversal tools - id: ECO-4.34 name: Profile differences from the engine-client breaker pillar: ECO-4 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/graph/parallel_engine.py - agent_utilities/knowledge_graph/core/engine_breaker.py - agent_utilities/knowledge_graph/core/engine_mcp_discovery.py - agent_utilities/mcp/child_resilience.py - agent_utilities/mcp/multiplexer.py - agent_utilities/observability/gateway_metrics.py doc: Profile differences from the engine-client breaker - id: ECO-4.35 name: Fleet HTTP Client Library pillar: ECO-4 status: live code_paths: - agent_utilities/http/__init__.py - agent_utilities/http/auth.py - agent_utilities/http/client.py - agent_utilities/http/pagination.py - agent_utilities/http/rate_limit.py - agent_utilities/http/redaction.py doc: Fleet HTTP Client Library - id: ECO-4.36 name: catalog-aware, collision-free prefix assignment pillar: ECO-4 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/mcp/eunomia_principal.py - agent_utilities/mcp/multiplexer.py doc: catalog-aware, collision-free prefix assignment - id: ECO-4.37 name: Surface Gateway Client SDK pillar: ECO-4 status: live code_paths: - agent_utilities/gateway_client/__init__.py - agent_utilities/gateway_client/client.py doc: Surface Gateway Client SDK - id: ECO-4.38 name: / ECO-4.42 pillar: ECO-4 status: live code_paths: - agent_utilities/ingestion/__init__.py - agent_utilities/ingestion/agent_sources/__init__.py - agent_utilities/ingestion/agent_sources/_jsonl.py - agent_utilities/ingestion/agent_sources/parsers/__init__.py - agent_utilities/ingestion/agent_sources/parsers/claude.py - agent_utilities/ingestion/agent_sources/parsers/codex.py - agent_utilities/ingestion/agent_sources/parsers/generic_jsonl.py - agent_utilities/ingestion/agent_sources/registry.py - agent_utilities/ingestion/collector.py doc: / ECO-4.42 - id: ECO-4.39 name: Usage analytics store. These pydantic models are the single pillar: ECO-4 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/usage/__init__.py - agent_utilities/usage/backend.py - agent_utilities/usage/backends/__init__.py - agent_utilities/usage/backends/duckdb_mirror.py - agent_utilities/usage/backends/postgres.py - agent_utilities/usage/backends/sql_base.py - agent_utilities/usage/backends/sqlite_fts.py - agent_utilities/usage/cost.py - agent_utilities/usage/models.py - agent_utilities/usage/recorder.py - agent_utilities/usage/schema.py - agent_utilities/usage/service.py doc: Usage analytics store. These pydantic models are the single - id: ECO-4.40 name: single source of token cost for the whole stack, replacing pillar: ECO-4 status: live code_paths: - agent_utilities/models/usage.py - agent_utilities/pricing/__init__.py - agent_utilities/pricing/catalog.py - agent_utilities/pricing/fallback.py - agent_utilities/pricing/litellm.py - agent_utilities/pricing/normalize.py - agent_utilities/pricing/store.py doc: single source of token cost for the whole stack, replacing - id: ECO-4.41 name: usage/cost/observability surface for all 3 UIs pillar: ECO-4 status: live code_paths: - agent_utilities/gateway/usage_api.py - agent_utilities/gateway_client/client.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/mcp/tools/write_ingest_tools.py - agent_utilities/server/app.py doc: usage/cost/observability surface for all 3 UIs - id: ECO-4.42 name: clients parse local logs and POST normalized pillar: ECO-4 status: live code_paths: - agent_utilities/gateway/usage_api.py - agent_utilities/mcp/tools/write_ingest_tools.py doc: clients parse local logs and POST normalized - id: ECO-4.43 name: git issue/PR -> SWE task resolver + webhook ingress pillar: ECO-4 status: live code_paths: - agent_utilities/integrations/git_resolver.py - agent_utilities/server/app.py - agent_utilities/server/routers/enhanced.py - agent_utilities/server/routers/git_webhooks.py doc: git issue/PR -> SWE task resolver + webhook ingress - id: ECO-4.44 name: Optional browser-agent tier for the developer-workspace runtime pillar: ECO-4 status: live code_paths: - agent_utilities/runtime/browser_tier.py - agent_utilities/runtime/events.py doc: Optional browser-agent tier for the developer-workspace runtime - id: ECO-4.45 name: bind tools from declared capability intents pillar: ECO-4 status: live code_paths: - agent_utilities/agent/capability_resolver.py - agent_utilities/agent/factory.py doc: bind tools from declared capability intents - id: ECO-4.46 name: ───────────────── pillar: ECO-4 status: live code_paths: - agent_utilities/protocols/source_connectors/connectors/mcp_tool.py doc: ───────────────── - id: ECO-4.47 name: Unattended overnight loop-driver + morning-summary write-back pillar: ECO-4 status: live code_paths: - agent_utilities/claude_harness/__init__.py - agent_utilities/claude_harness/overnight_runner.py - agent_utilities/cli/__init__.py doc: Unattended overnight loop-driver + morning-summary write-back - id: EE-033 name: closes the priors→weights loop pillar: EE-033 status: live code_paths: - agent_utilities/domains/finance/signal_fusion.py - agent_utilities/knowledge_graph/orchestration/engine_finance.py - agent_utilities/models/domains/finance.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/schema_packs/finance.py doc: closes the priors→weights loop - id: EE-034 name: the expert agent writes one per decision; a nightly distill pillar: EE-034 status: live code_paths: - agent_utilities/models/domains/finance.py doc: the expert agent writes one per decision; a nightly distill - id: EE-036 name: concrete subclasses pillar: EE-036 status: live code_paths: - agent_utilities/knowledge_graph/distillation/trading_curator.py - agent_utilities/models/domains/finance.py - agent_utilities/models/schema_packs/finance.py doc: concrete subclasses - id: EE-037 name: microstructure, trading, pricing pillar: EE-037 status: live code_paths: - agent_utilities/automation/research_pipeline.py doc: microstructure, trading, pricing - id: EE-039 name: EE-039 pillar: EE-039 status: live code_paths: - agent_utilities/graph/trading_team_seed.py doc: EE-039 - id: EG-009 name: then by its node pillar: EG-009 status: live code_paths: - agent_utilities/knowledge_graph/standardization/standards.py doc: then by its node - id: KG-1.0 name: Centralized KG Coordination Protocol pillar: KG-1 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/mcp/kg_coordinator.py doc: Centralized KG Coordination Protocol - id: KG-2.0 name: Provides git-like transactional mutation for KG evolution pillar: KG-2 status: live code_paths: - agent_utilities/automation/research_pipeline.py - agent_utilities/core/cognitive_scheduler.py - agent_utilities/core/registry/kg_adapter.py - agent_utilities/core/registry/package_adapter.py - agent_utilities/domains/government/__init__.py - agent_utilities/domains/government/models.py - agent_utilities/graph/reward_decomposition.py - agent_utilities/graph/workspace_attention.py - agent_utilities/harness/variant_pool.py - agent_utilities/knowledge_graph/__main__.py - agent_utilities/knowledge_graph/adaptation/research_artifacts.py - agent_utilities/knowledge_graph/backends/contrib/falkordb_backend.py - agent_utilities/knowledge_graph/backends/contrib/ladybug_backend.py - agent_utilities/knowledge_graph/backends/contrib/neo4j_backend.py - agent_utilities/knowledge_graph/backends/epistemic_graph_backend.py - agent_utilities/knowledge_graph/core/codemaps.py - agent_utilities/knowledge_graph/core/context_builder.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/graph_compute.py - agent_utilities/knowledge_graph/core/kg_versioning.py - agent_utilities/knowledge_graph/core/maintainer.py - agent_utilities/knowledge_graph/core/ogm.py - agent_utilities/knowledge_graph/core/source_resolver.py - agent_utilities/knowledge_graph/kb/extractor.py - agent_utilities/knowledge_graph/kb/ingestion.py - agent_utilities/knowledge_graph/kb/parser.py - agent_utilities/knowledge_graph/maintenance/document_cleanup.py - agent_utilities/knowledge_graph/memory/strategies/heavy_thinking_cache.py - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/knowledge_graph/pipeline/document_deletion.py - agent_utilities/knowledge_graph/pipeline/document_ingestion.py - agent_utilities/knowledge_graph/pipeline/document_update.py - agent_utilities/knowledge_graph/pipeline/phases/parse.py - agent_utilities/knowledge_graph/pipeline/phases/resolve.py - agent_utilities/knowledge_graph/pipeline/phases/sync.py - agent_utilities/knowledge_graph/pipeline/phases/workspace_sync.py - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/knowledge_graph/retrieval/memory_retriever.py - agent_utilities/knowledge_graph/self_model.py - agent_utilities/models/codemap.py - agent_utilities/models/knowledge_base.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/schema_definition.py - agent_utilities/tools/kg_share_tools.py - agent_utilities/tools/knowledge_tools.py doc: Provides git-like transactional mutation for KG evolution - id: KG-2.1 name: routes skill evolution through the single graph-native pillar: KG-2 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/capabilities/eviction.py - agent_utilities/core/chat_persistence.py - agent_utilities/core/paths.py - agent_utilities/graph/_router_impl.py - agent_utilities/graph/executor.py - agent_utilities/graph/parallel_engine.py - agent_utilities/graph/routing/enrichers/capability_designation.py - agent_utilities/graph/routing/enrichers/self_model.py - agent_utilities/graph/routing/strategies/fast_path.py - agent_utilities/graph/routing/strategies/workflow_context.py - agent_utilities/graph/subagent_patterns.py - agent_utilities/graph/verification.py - agent_utilities/graph/workspace_attention.py - agent_utilities/harness/__init__.py - agent_utilities/harness/agentic_evolution_engine.py - agent_utilities/harness/evolving_memory.py - agent_utilities/knowledge_graph/core/agents_md.py - agent_utilities/knowledge_graph/core/conversation_ingestion.py - agent_utilities/knowledge_graph/core/engine_memory.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/model_display.py - agent_utilities/knowledge_graph/ingestion/engine.py - agent_utilities/knowledge_graph/memory/__init__.py - agent_utilities/knowledge_graph/memory/agent_context.py - agent_utilities/knowledge_graph/memory/cli.py - agent_utilities/knowledge_graph/memory/memory_engine.py - agent_utilities/knowledge_graph/memory/observer.py - agent_utilities/knowledge_graph/memory/optimization_engine.py - agent_utilities/knowledge_graph/orchestration/engine_federation.py - agent_utilities/knowledge_graph/pipeline/phases/external_graphs.py - agent_utilities/knowledge_graph/retrieval/budget.py - agent_utilities/knowledge_graph/retrieval/chat_search.py - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/knowledge_graph/retrieval/memory_retriever.py - agent_utilities/messaging/kg_ingest.py - agent_utilities/messaging/router.py - agent_utilities/models/company.py - agent_utilities/models/knowledge_graph.py - agent_utilities/prompting/builder.py doc: routes skill evolution through the single graph-native - id: KG-2.2 name: the self-bootstrapping ontology agent applied to ingest pillar: KG-2 status: live code_paths: - agent_utilities/automation/research_pipeline.py - agent_utilities/ecosystem/governance_agent.py - agent_utilities/knowledge_graph/backends/owl/stardog_backend.py - agent_utilities/knowledge_graph/core/engine_ingestion.py - agent_utilities/knowledge_graph/core/graph_compute.py - agent_utilities/knowledge_graph/core/inference_engine.py - agent_utilities/knowledge_graph/core/model_display.py - agent_utilities/knowledge_graph/core/ontology_bootstrap.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/core/semantic_subsumption.py - agent_utilities/knowledge_graph/distillation/__init__.py - agent_utilities/knowledge_graph/distillation/deduplicator.py - agent_utilities/knowledge_graph/distillation/distillation_engine.py - agent_utilities/knowledge_graph/distillation/lsh_index.py - agent_utilities/knowledge_graph/dspy_kg_bridge.py - agent_utilities/knowledge_graph/id_management/ontological_identifier.py - agent_utilities/knowledge_graph/kb/entity_claim_extractor.py - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/knowledge_graph/pipeline/phases/owl_reasoning.py - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/knowledge_graph/security/policy_ingestor.py - agent_utilities/knowledge_graph/security/rule_ingestor.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/models/imodel.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/schema_pack.py - agent_utilities/models/schema_packs/__init__.py - agent_utilities/policies/__init__.py - agent_utilities/prompting/dspy_compiler.py - agent_utilities/prompting/provider_adapter.py - agent_utilities/prompting/structured.py doc: the self-bootstrapping ontology agent applied to ingest - id: KG-2.3 name: Lazy embedding model — defer HTTP connection to first use pillar: KG-2 status: live code_paths: - agent_utilities/core/embedding_utilities.py - agent_utilities/graph/hierarchical_planner.py - agent_utilities/harness/citation_tracker.py - agent_utilities/knowledge_graph/adaptation/experience_alignment.py - agent_utilities/knowledge_graph/core/fingerprint.py - agent_utilities/knowledge_graph/memory/optimization_engine.py - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/knowledge_graph/pipeline/phases/embedding.py - agent_utilities/knowledge_graph/pipeline/phases/validate.py - agent_utilities/knowledge_graph/retrieval/evaluation_corpus.py - agent_utilities/knowledge_graph/retrieval/hard_negative_miner.py - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/knowledge_graph/retrieval/retrieval_quality.py - agent_utilities/knowledge_graph/retrieval/semantic_retrieval_engine.py - agent_utilities/knowledge_graph/security/cognitive_trap_defense.py - agent_utilities/knowledge_graph/security/graph_validator.py - agent_utilities/mcp/tools/query_tools.py - agent_utilities/messaging/kg_ingest.py - agent_utilities/models/knowledge_graph.py doc: Lazy embedding model — defer HTTP connection to first use - id: KG-2.4 name: Compute positional interaction encoding for structural generalization pillar: KG-2 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/graph/verification.py - agent_utilities/knowledge_graph/adaptation/trace_distiller.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/hypergraph.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/core/synergy_engine.py - agent_utilities/knowledge_graph/memory/strategies/heavy_thinking_cache.py - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/models/imodel.py - agent_utilities/models/knowledge_graph.py doc: Compute positional interaction encoding for structural generalization - id: KG-2.5 name: /2.15/2.34/2.35 — Topological Analysis Engine pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/analogy_engine.py - agent_utilities/knowledge_graph/core/blast_radius.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/hierarchical_retrieval.py - agent_utilities/knowledge_graph/core/model_display.py - agent_utilities/knowledge_graph/core/spectral_navigator.py - agent_utilities/knowledge_graph/core/topological_analysis_engine.py - agent_utilities/knowledge_graph/core/topological_partition.py - agent_utilities/knowledge_graph/memory/rlm_memory.py - agent_utilities/knowledge_graph/orchestration/voi_budget_controller.py - agent_utilities/knowledge_graph/retrieval/single_shot_sira.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/schema_definition.py doc: /2.15/2.34/2.35 — Topological Analysis Engine - id: KG-2.6 name: Generates actionable LLM artifacts from KG-ingested research pillar: KG-2 status: live code_paths: - agent_utilities/automation/__init__.py - agent_utilities/automation/research_pipeline.py - agent_utilities/core/checkpoint/manager.py - agent_utilities/domains/__init__.py - agent_utilities/domains/finance/alpha_factors.py - agent_utilities/domains/finance/banking.py - agent_utilities/domains/finance/banking_models.py - agent_utilities/domains/finance/copy_trade.py - agent_utilities/domains/finance/credit_quality.py - agent_utilities/domains/finance/cross_market_arb.py - agent_utilities/domains/finance/crypto_connector.py - agent_utilities/domains/finance/debate_engine.py - agent_utilities/domains/finance/evaluation.py - agent_utilities/domains/finance/exchange_bridge.py - agent_utilities/domains/finance/filing_diff.py - agent_utilities/domains/finance/forensic_screener.py - agent_utilities/domains/finance/insider_equilibrium.py - agent_utilities/domains/finance/investor_debate.py - agent_utilities/domains/finance/kronos_forecaster.py - agent_utilities/domains/finance/market_data.py - agent_utilities/domains/finance/market_feeds.py - agent_utilities/domains/finance/microstructure.py - agent_utilities/domains/finance/pattern_classifier.py - agent_utilities/domains/finance/payments.py - agent_utilities/domains/finance/portfolio_optimizer.py - agent_utilities/domains/finance/profit_attribution.py - agent_utilities/domains/finance/quant_ontology.py - agent_utilities/domains/finance/regime_detector.py - agent_utilities/domains/finance/research_autopilot.py - agent_utilities/domains/finance/risk_manager.py - agent_utilities/domains/finance/signal_fusion.py - agent_utilities/domains/finance/strategy_engine.py - agent_utilities/domains/finance/strategy_export.py - agent_utilities/domains/finance/strategy_sharing.py - agent_utilities/domains/finance/streaming.py - agent_utilities/domains/finance/trading_swarm.py - agent_utilities/domains/finance/versioned_orders.py - agent_utilities/domains/finance/visual_ta.py - agent_utilities/domains/hr/__init__.py - agent_utilities/domains/hr/models.py - agent_utilities/domains/hr/workforce_manager.py - agent_utilities/domains/law/__init__.py - agent_utilities/domains/law/models.py - agent_utilities/domains/medical/__init__.py - agent_utilities/domains/medical/models.py - agent_utilities/graph/state.py - agent_utilities/knowledge_graph/adaptation/research_artifacts.py - agent_utilities/knowledge_graph/backends/__init__.py - agent_utilities/knowledge_graph/backends/brain_guarded_backend.py - agent_utilities/knowledge_graph/backends/epistemic_graph_backend.py - agent_utilities/knowledge_graph/core/archimate_layer.py - agent_utilities/knowledge_graph/core/company_brain.py - agent_utilities/knowledge_graph/core/company_brain_runtime.py - agent_utilities/knowledge_graph/core/engine.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/formal_reasoning_core.py - agent_utilities/knowledge_graph/core/markov_regime.py - agent_utilities/knowledge_graph/core/model_display.py - agent_utilities/knowledge_graph/core/ontological_team_sharing.py - agent_utilities/knowledge_graph/core/ontology_loader.py - agent_utilities/knowledge_graph/core/ontology_publisher.py - agent_utilities/knowledge_graph/core/optimal_execution.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/core/secured_reads.py - agent_utilities/knowledge_graph/core/shacl_validator.py - agent_utilities/knowledge_graph/core/source_resolver.py - agent_utilities/knowledge_graph/core/sparql_http.py - agent_utilities/knowledge_graph/kb/knowledge_classifier.py - agent_utilities/knowledge_graph/kb/x_workflows.py - agent_utilities/knowledge_graph/maintenance/owl_closure.py - agent_utilities/knowledge_graph/memory/optimization_engine.py - agent_utilities/knowledge_graph/orchestration/engine_enterprise.py - agent_utilities/knowledge_graph/orchestration/engine_finance.py - agent_utilities/knowledge_graph/orchestration/engine_ml_rlm.py - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/knowledge_graph/orchestration/research_orchestrator.py - agent_utilities/knowledge_graph/orchestration/research_subagent.py - agent_utilities/knowledge_graph/research/research_intelligence_engine.py - agent_utilities/knowledge_graph/retrieval/embedding_diagnostics.py - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/knowledge_graph/retrieval/latent_topology_rag.py - agent_utilities/knowledge_graph/retrieval/reasoning_reranker.py - agent_utilities/knowledge_graph/retrieval/semantic_retrieval_engine.py - agent_utilities/knowledge_graph/streams/__init__.py - agent_utilities/knowledge_graph/streams/kafka_adapter.py - agent_utilities/knowledge_graph/streams/nats_adapter.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/models/company.py - agent_utilities/models/company_brain.py - agent_utilities/models/domains/finance.py - agent_utilities/models/imodel.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/knowledge_pack.py - agent_utilities/models/schema_pack.py - agent_utilities/models/schema_packs/finance.py - agent_utilities/sdd/watcher.py - agent_utilities/security/brain_context.py - agent_utilities/security/threat_defense_engine.py doc: Generates actionable LLM artifacts from KG-ingested research - id: KG-2.7 name: / KG-2.10 / KG-2.78 — research assimilation + orchestration synthesis pillar: KG-2 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/api/sparql_server.py - agent_utilities/core/background_throttle.py - agent_utilities/core/config.py - agent_utilities/core/workspace_config.py - agent_utilities/gateway/graph_api.py - agent_utilities/graph/routing/strategies/query_tier.py - agent_utilities/knowledge_graph/adaptation/topic_resolver.py - agent_utilities/knowledge_graph/assimilation/__init__.py - agent_utilities/knowledge_graph/assimilation/breadth_ingest.py - agent_utilities/knowledge_graph/assimilation/dedup.py - agent_utilities/knowledge_graph/assimilation/gap_analysis.py - agent_utilities/knowledge_graph/assimilation/ingest.py - agent_utilities/knowledge_graph/assimilation/ledger.py - agent_utilities/knowledge_graph/assimilation/pilot.py - agent_utilities/knowledge_graph/assimilation/plan_synthesis.py - agent_utilities/knowledge_graph/assimilation/synergy.py - agent_utilities/knowledge_graph/backends/age_backend.py - agent_utilities/knowledge_graph/backends/base.py - agent_utilities/knowledge_graph/backends/contrib/ladybug_backend.py - agent_utilities/knowledge_graph/backends/cypher_transpiler.py - agent_utilities/knowledge_graph/backends/epistemic_graph_backend.py - agent_utilities/knowledge_graph/backends/postgresql_backend.py - agent_utilities/knowledge_graph/backends/sparql/base.py - agent_utilities/knowledge_graph/backends/sparql/jena_fuseki_backend.py - agent_utilities/knowledge_graph/backends/sparql/source_partition.py - agent_utilities/knowledge_graph/backends/sparql/stardog_backend.py - agent_utilities/knowledge_graph/backends/tiered_backend.py - agent_utilities/knowledge_graph/core/company_brain.py - agent_utilities/knowledge_graph/core/engine.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/event_backend.py - agent_utilities/knowledge_graph/core/graph_compute.py - agent_utilities/knowledge_graph/core/hydration.py - agent_utilities/knowledge_graph/core/kafka_graph_sync.py - agent_utilities/knowledge_graph/core/kg_versioning.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/core/source_sync.py - agent_utilities/knowledge_graph/core/subgraph_checkout.py - agent_utilities/knowledge_graph/distillation/skill_graph_distiller.py - agent_utilities/knowledge_graph/distillation/skill_graph_pipeline.py - agent_utilities/knowledge_graph/distillation/skill_graph_schema.py - agent_utilities/knowledge_graph/enrichment/execute.py - agent_utilities/knowledge_graph/enrichment/models.py - agent_utilities/knowledge_graph/ingestion/__init__.py - agent_utilities/knowledge_graph/ingestion/__main__.py - agent_utilities/knowledge_graph/ingestion/engine.py - agent_utilities/knowledge_graph/ingestion/paper_links.py - agent_utilities/knowledge_graph/ingestion/research_acquisition.py - agent_utilities/knowledge_graph/ingestion/web_fetch.py - agent_utilities/knowledge_graph/integrations/stardog_sync.py - agent_utilities/knowledge_graph/memory/agent_context.py - agent_utilities/knowledge_graph/ontology/interfaces.py - agent_utilities/knowledge_graph/pipeline/__init__.py - agent_utilities/knowledge_graph/pipeline/phases/__init__.py - agent_utilities/knowledge_graph/research/loop_controller.py - agent_utilities/knowledge_graph/research/search.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/mcp/tools/write_ingest_tools.py - agent_utilities/models/knowledge_graph.py - agent_utilities/server/app.py doc: / KG-2.10 / KG-2.78 — research assimilation + orchestration synthesis - id: KG-2.8 name: / OS-5.9 — the KG runs exactly ONE consolidated background daemon pillar: KG-2 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/gateway/api.py - agent_utilities/gateway/daemon.py - agent_utilities/knowledge_graph/adaptation/feedback.py - agent_utilities/knowledge_graph/assimilation/breadth_ingest.py - agent_utilities/knowledge_graph/backends/__init__.py - agent_utilities/knowledge_graph/backends/epistemic_graph_backend.py - agent_utilities/knowledge_graph/backends/owl/owlready2_backend.py - agent_utilities/knowledge_graph/core/conversation_ingestion.py - agent_utilities/knowledge_graph/core/engine_lock.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/graph_compute.py - agent_utilities/knowledge_graph/core/host_lock.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/enrichment/__init__.py - agent_utilities/knowledge_graph/enrichment/__main__.py - agent_utilities/knowledge_graph/enrichment/cards.py - agent_utilities/knowledge_graph/enrichment/classify.py - agent_utilities/knowledge_graph/enrichment/distill.py - agent_utilities/knowledge_graph/enrichment/extractors/__init__.py - agent_utilities/knowledge_graph/enrichment/extractors/code_test.py - agent_utilities/knowledge_graph/enrichment/extractors/document.py - agent_utilities/knowledge_graph/enrichment/extractors/text.py - agent_utilities/knowledge_graph/enrichment/features.py - agent_utilities/knowledge_graph/enrichment/models.py - agent_utilities/knowledge_graph/enrichment/patterns.py - agent_utilities/knowledge_graph/enrichment/pipeline.py - agent_utilities/knowledge_graph/enrichment/query.py - agent_utilities/knowledge_graph/enrichment/realizes.py - agent_utilities/knowledge_graph/enrichment/semantic.py - agent_utilities/knowledge_graph/enrichment/writeback/__init__.py - agent_utilities/knowledge_graph/enrichment/writeback/core.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/capability.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/process.py - agent_utilities/knowledge_graph/facade.py - agent_utilities/knowledge_graph/ingest_worker.py - agent_utilities/knowledge_graph/ingestion/engine.py - agent_utilities/knowledge_graph/ingestion/manifest.py - agent_utilities/knowledge_graph/pipeline/phases/shacl_gate.py - agent_utilities/knowledge_graph/research/code_synthesis.py - agent_utilities/knowledge_graph/retrieval/governance_rules.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/mcp/tools/write_ingest_tools.py - agent_utilities/models/knowledge_graph.py - agent_utilities/orchestration/agent_dispatch_worker.py - agent_utilities/protocols/source_connectors/checkpoint.py - agent_utilities/sdd/watcher.py - agent_utilities/server/routers/commands.py doc: / OS-5.9 — the KG runs exactly ONE consolidated background daemon - id: KG-2.8r name: to turn each project pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/gitlab_indexer.py - agent_utilities/knowledge_graph/core/graph_compute.py - agent_utilities/knowledge_graph/core/source_sync.py doc: to turn each project - id: KG-2.9 name: which speaks **PostgreSQL, MySQL/MariaDB, MS SQL Server, Oracle pillar: KG-2 status: live code_paths: - agent_utilities/ecosystem/ea_clients.py - agent_utilities/gateway/ontology_api.py - agent_utilities/knowledge_graph/core/engine_ingestion.py - agent_utilities/knowledge_graph/core/hydration.py - agent_utilities/knowledge_graph/core/materialization.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/core/source_sync.py - agent_utilities/knowledge_graph/enrichment/extractors/ansible.py - agent_utilities/knowledge_graph/enrichment/extractors/archer.py - agent_utilities/knowledge_graph/enrichment/extractors/archimate.py - agent_utilities/knowledge_graph/enrichment/extractors/aris.py - agent_utilities/knowledge_graph/enrichment/extractors/caddy.py - agent_utilities/knowledge_graph/enrichment/extractors/camunda.py - agent_utilities/knowledge_graph/enrichment/extractors/egeria.py - agent_utilities/knowledge_graph/enrichment/extractors/emerald.py - agent_utilities/knowledge_graph/enrichment/extractors/erpnext.py - agent_utilities/knowledge_graph/enrichment/extractors/grafana.py - agent_utilities/knowledge_graph/enrichment/extractors/home_assistant.py - agent_utilities/knowledge_graph/enrichment/extractors/infra.py - agent_utilities/knowledge_graph/enrichment/extractors/kafka.py - agent_utilities/knowledge_graph/enrichment/extractors/keycloak.py - agent_utilities/knowledge_graph/enrichment/extractors/leanix.py - agent_utilities/knowledge_graph/enrichment/extractors/lgtm.py - agent_utilities/knowledge_graph/enrichment/extractors/mealie.py - agent_utilities/knowledge_graph/enrichment/extractors/microsoft.py - agent_utilities/knowledge_graph/enrichment/extractors/nextcloud.py - agent_utilities/knowledge_graph/enrichment/extractors/odoo.py - agent_utilities/knowledge_graph/enrichment/extractors/okta.py - agent_utilities/knowledge_graph/enrichment/extractors/portainer.py - agent_utilities/knowledge_graph/enrichment/extractors/salesforce.py - agent_utilities/knowledge_graph/enrichment/extractors/technitium_dns.py - agent_utilities/knowledge_graph/enrichment/extractors/twenty.py - agent_utilities/knowledge_graph/enrichment/extractors/uptime_kuma.py - agent_utilities/knowledge_graph/enrichment/extractors/wger.py - agent_utilities/knowledge_graph/enrichment/materialize.py - agent_utilities/knowledge_graph/enrichment/models.py - agent_utilities/knowledge_graph/enrichment/provenance.py - agent_utilities/knowledge_graph/enrichment/registry.py - agent_utilities/knowledge_graph/enrichment/source_adapters.py - agent_utilities/knowledge_graph/enrichment/writeback/approval.py - agent_utilities/knowledge_graph/enrichment/writeback/findings.py - agent_utilities/knowledge_graph/enrichment/writeback/inventory.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/ansible.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/archimate.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/egeria.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/emerald.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/erpnext.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/home_assistant.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/identity.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/issue_tracker.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/leanix.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/legal.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/mealie.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/nextcloud.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/ops.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/salesforce.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/servicenow.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/twenty.py - agent_utilities/knowledge_graph/enrichment/writeback/sinks/wger.py - agent_utilities/knowledge_graph/enrichment/writeback/spec_link.py - agent_utilities/knowledge_graph/infra/__init__.py - agent_utilities/knowledge_graph/infra/inventory_collector.py - agent_utilities/knowledge_graph/infra/placement_optimizer.py - agent_utilities/knowledge_graph/ontology/leanix_metamodel.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/mcp/tools/state_tools.py - agent_utilities/mcp/tools/write_ingest_tools.py - agent_utilities/models/sdd.py - agent_utilities/protocols/source_connectors/connectors/database.py - agent_utilities/protocols/universal_connector.py - agent_utilities/tools/db_tools.py doc: which speaks **PostgreSQL, MySQL/MariaDB, MS SQL Server, Oracle - id: KG-2.9g name: Declared columns so schema-backed Kuzu can pillar: KG-2 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/backends/contrib/ladybug_backend.py - agent_utilities/knowledge_graph/backends/epistemic_graph_backend.py - agent_utilities/knowledge_graph/backends/tiered_backend.py - agent_utilities/knowledge_graph/core/gitlab_indexer.py - agent_utilities/knowledge_graph/core/source_sync.py - agent_utilities/mcp/tools/query_tools.py - agent_utilities/models/schema_definition.py - agent_utilities/protocols/source_connectors/connectors/mcp_tool.py doc: Declared columns so schema-backed Kuzu can - id: KG-2.10 name: S4c pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/enrichment/execute.py - agent_utilities/knowledge_graph/enrichment/extractors/a2a.py - agent_utilities/knowledge_graph/enrichment/extractors/config_models.py - agent_utilities/knowledge_graph/enrichment/pydantic_graph.py - agent_utilities/knowledge_graph/enrichment/synthesize.py doc: S4c - id: KG-2.11 name: optional ISO-8601 instant. When set, rows are filtered to pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/bitemporal.py - agent_utilities/knowledge_graph/core/engine.py - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/mcp/tools/query_tools.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/schema_pack.py doc: optional ISO-8601 instant. When set, rows are filtered to - id: KG-2.12 name: run a self-correcting second retrieval pass at the deep threshold when the quality gate fails pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/knowledge_graph/retrieval/direct_corpus.py - agent_utilities/knowledge_graph/retrieval/executable_rag.py - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/knowledge_graph/retrieval/hyde_planner.py - agent_utilities/mcp/tools/query_tools.py doc: run a self-correcting second retrieval pass at the deep threshold when the quality gate fails - id: KG-2.13 name: targeted ADD/UPDATE/DELETE learner pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/memory/cli.py - agent_utilities/knowledge_graph/memory/learning_engine.py doc: targeted ADD/UPDATE/DELETE learner - id: KG-2.14 name: Ground-Truth preamble declaring authoritative sources up front pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/memory/memory_engine.py doc: Ground-Truth preamble declaring authoritative sources up front - id: KG-2.15 name: social-closer / triviality gate pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/knowledge_graph/retrieval/hyde_planner.py doc: social-closer / triviality gate - id: KG-2.16 name: KG-2.16 pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/graph_compute.py - agent_utilities/knowledge_graph/enrichment/extractors/code_test.py - agent_utilities/knowledge_graph/enrichment/features.py - agent_utilities/knowledge_graph/enrichment/pipeline.py - agent_utilities/knowledge_graph/ingestion/engine.py - agent_utilities/knowledge_graph/orchestration/engine_query.py doc: KG-2.16 - id: KG-2.17 name: decay scanner + semantic merge pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/memory/cli.py - agent_utilities/knowledge_graph/memory/hygiene.py - agent_utilities/knowledge_graph/memory/memory_engine.py doc: decay scanner + semantic merge - id: KG-2.18 name: record that these memories were RECALLED, on the live retrieval path pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/knowledge_graph/retrieval/retrieval_quality.py doc: record that these memories were RECALLED, on the live retrieval path - id: KG-2.19 name: delta-skip continuous ingest of a self-curating wiki dir pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/ingestion/batch_orchestrator.py - agent_utilities/knowledge_graph/ingestion/wiki_curator.py - agent_utilities/mcp/tools/write_ingest_tools.py doc: delta-skip continuous ingest of a self-curating wiki dir - id: KG-2.20 name: live Memento block-compress-evict sawtooth pillar: KG-2 status: live code_paths: - agent_utilities/agent/factory.py - agent_utilities/capabilities/memento.py - agent_utilities/knowledge_graph/memory/agent_context.py - agent_utilities/knowledge_graph/memory/memento_compressor.py doc: live Memento block-compress-evict sawtooth - id: KG-2.20g name: forensic accounting kernels pillar: KG-2 status: live code_paths: - agent_utilities/domains/finance/forensic_screener.py doc: forensic accounting kernels - id: KG-2.21 name: Working Set Eviction & Memory Management pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/working_set_manager.py doc: Working Set Eviction & Memory Management - id: KG-2.22 name: Pack-Driven Retrieval Signals pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/knowledge_graph/retrieval/autocut.py - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/mcp/tools/query_tools.py - agent_utilities/models/schema_pack.py - agent_utilities/models/schema_packs/research.py doc: Pack-Driven Retrieval Signals - id: KG-2.23 name: Rust-Accelerated Reasoning pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py doc: Rust-Accelerated Reasoning - id: KG-2.24 name: Live Refreshable Artifact models + bounded-JSON + safe interpolation pillar: KG-2 status: live code_paths: - agent_utilities/gateway/artifacts_api.py - agent_utilities/knowledge_graph/live_artifacts/__init__.py - agent_utilities/knowledge_graph/live_artifacts/kg_source.py - agent_utilities/knowledge_graph/live_artifacts/models.py - agent_utilities/knowledge_graph/live_artifacts/refresh.py - agent_utilities/knowledge_graph/live_artifacts/store.py - agent_utilities/server/app.py doc: Live Refreshable Artifact models + bounded-JSON + safe interpolation - id: KG-2.25 name: the default registry is populated at import, not an empty pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/actions/__init__.py - agent_utilities/knowledge_graph/actions/builtins.py - agent_utilities/knowledge_graph/actions/executor.py - agent_utilities/knowledge_graph/actions/models.py - agent_utilities/knowledge_graph/actions/registry.py - agent_utilities/knowledge_graph/backends/owl/owlready2_backend.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/core/shacl_validator.py - agent_utilities/knowledge_graph/ontology/functions/__init__.py - agent_utilities/knowledge_graph/ontology/functions/runtime.py doc: the default registry is populated at import, not an empty - id: KG-2.26 name: the Palantir *many-to-many link backed by an object type* pillar: KG-2 status: live code_paths: - agent_utilities/domains/finance/trade_journal.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/links.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/mcp/tools/state_tools.py doc: the Palantir *many-to-many link backed by an object type* - id: KG-2.27 name: Agent Calibration and Reputation Tracking pillar: KG-2 status: live code_paths: - agent_utilities/domains/finance/calibration_tracker.py - agent_utilities/domains/finance/trading_swarm.py doc: Agent Calibration and Reputation Tracking - id: KG-2.28 name: Persona Decision-Heuristic Enrichment pillar: KG-2 status: live code_paths: - agent_utilities/domains/finance/persona_heuristics.py doc: Persona Decision-Heuristic Enrichment - id: KG-2.29 name: Sentiment Fusion Signals pillar: KG-2 status: live code_paths: - agent_utilities/domains/finance/sentiment_fusion.py doc: Sentiment Fusion Signals - id: KG-2.30 name: Geopolitical Risk Scoring pillar: KG-2 status: live code_paths: - agent_utilities/domains/finance/geopolitical_risk.py doc: Geopolitical Risk Scoring - id: KG-2.31 name: Dividend Sustainability & Credit/Fixed-Income Quality pillar: KG-2 status: live code_paths: - agent_utilities/domains/finance/credit_quality.py doc: Dividend Sustainability & Credit/Fixed-Income Quality - id: KG-2.32 name: Multi-Market Composite Backtester pillar: KG-2 status: live code_paths: - agent_utilities/domains/finance/composite_backtest.py doc: Multi-Market Composite Backtester - id: KG-2.33 name: Zero-LLM Pack-Driven Link Inference pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/kb/entity_claim_extractor.py - agent_utilities/knowledge_graph/kb/link_inference.py - agent_utilities/models/schema_pack.py - agent_utilities/models/schema_packs/research.py doc: Zero-LLM Pack-Driven Link Inference - id: KG-2.34 name: Relational-Intent Retrieval pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/retrieval/hybrid_retriever.py - agent_utilities/knowledge_graph/retrieval/relational_intent.py - agent_utilities/models/schema_pack.py - agent_utilities/models/schema_packs/research.py doc: Relational-Intent Retrieval - id: KG-2.35 name: Schema-Pack Lifecycle and Audit pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/engine.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/models/schema_pack_audit.py - agent_utilities/models/schema_pack_loader.py doc: Schema-Pack Lifecycle and Audit - id: KG-2.36 name: reuses the pack-driven OWL closure notion. The pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/ontology/links.py - agent_utilities/models/schema_pack.py - agent_utilities/models/schema_packs/research.py doc: reuses the pack-driven OWL closure notion. The - id: KG-2.37 name: Research-State Domain Pack pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/schema_packs/research.py doc: Research-State Domain Pack - id: KG-2.38 name: the function the A3 functions runtime / object queries call pillar: KG-2 status: live code_paths: - agent_utilities/gateway/ontology_api.py - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/interfaces.py - agent_utilities/knowledge_graph/ontology/object_set.py - agent_utilities/knowledge_graph/standardization/standards.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/mcp/tools/state_tools.py doc: the function the A3 functions runtime / object queries call - id: KG-2.39 name: the live consumer hook. Writes the combined SHACL document pillar: KG-2 status: live code_paths: - agent_utilities/gateway/ontology_api.py - agent_utilities/knowledge_graph/core/shacl_validator.py - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/value_types.py - agent_utilities/knowledge_graph/pipeline/phases/shacl_gate.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/mcp/tools/state_tools.py doc: the live consumer hook. Writes the combined SHACL document - id: KG-2.40 name: import-populated registry + a default engine bound to it and pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/derived_properties.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/mcp/tools/state_tools.py doc: import-populated registry + a default engine bound to it and - id: KG-2.41 name: a default runtime bound to the import-populated registry, so pillar: KG-2 status: live code_paths: - agent_utilities/gateway/ontology_api.py - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/functions/__init__.py - agent_utilities/knowledge_graph/ontology/functions/objects.py - agent_utilities/knowledge_graph/ontology/functions/registry.py - agent_utilities/knowledge_graph/ontology/functions/runtime.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/mcp/tools/state_tools.py doc: a default runtime bound to the import-populated registry, so - id: KG-2.42 name: Palantir submission-criteria gate. An action with no pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/actions/builtins.py - agent_utilities/knowledge_graph/actions/dispatch.py - agent_utilities/knowledge_graph/actions/effects.py - agent_utilities/knowledge_graph/actions/executor.py - agent_utilities/knowledge_graph/actions/fleet_writeback.py - agent_utilities/knowledge_graph/actions/models.py - agent_utilities/protocols/source_connectors/connectors/mcp_tool.py doc: Palantir submission-criteria gate. An action with no - id: KG-2.43 name: the write-back leg of the edit ledger. Sinks are registered pillar: KG-2 status: live code_paths: - agent_utilities/gateway/ontology_api.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/edits/__init__.py - agent_utilities/knowledge_graph/ontology/edits/ledger.py - agent_utilities/knowledge_graph/ontology/edits/revert.py - agent_utilities/knowledge_graph/ontology/edits/writeback.py - agent_utilities/mcp/tools/ontology_tools.py doc: the write-back leg of the edit ledger. Sinks are registered - id: KG-2.44 name: batch/incremental sync of the live search index from source nodes, report staleness, or reindex stale objects pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/facade.py - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/indexing/__init__.py - agent_utilities/knowledge_graph/ontology/indexing/funnel.py - agent_utilities/knowledge_graph/ontology/indexing/staleness.py - agent_utilities/mcp/tools/ontology_tools.py doc: batch/incremental sync of the live search index from source nodes, report staleness, or reindex stale objects - id: KG-2.44b name: Optional; populated from the pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/retrieval/capability_index.py doc: Optional; populated from the - id: KG-2.45 name: pivot across a link type. Computes the 1-hop related set pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/object_set.py - agent_utilities/mcp/tools/ontology_tools.py doc: pivot across a link type. Computes the 1-hop related set - id: KG-2.46 name: redact an object, materialize a restricted view, or attach a mandatory marking. Actor is resolved from the ambient context — never from caller-supplied clearance pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/facade.py - agent_utilities/knowledge_graph/ontology/permissioning.py - agent_utilities/mcp/tools/ontology_tools.py doc: redact an object, materialize a restricted view, or attach a mandatory marking. Actor is resolved from the ambient context — never from caller-supplied clearance - id: KG-2.47 name: bridges the ontology type vocabulary onto the exact pillar: KG-2 status: live code_paths: - agent_utilities/gateway/ontology_api.py - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/property_types.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/mcp/tools/state_tools.py doc: bridges the ontology type vocabulary onto the exact - id: KG-2.48 name: the one-call form the ingestion engine / MCP wrapper invokes pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/facade.py - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/document_processing.py - agent_utilities/mcp/tools/ontology_tools.py doc: the one-call form the ingestion engine / MCP wrapper invokes - id: KG-2.49 name: populated at import with real built-in standards, never empty pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/research/loop_controller.py - agent_utilities/knowledge_graph/standardization/__init__.py - agent_utilities/knowledge_graph/standardization/cohorts.py - agent_utilities/knowledge_graph/standardization/consolidation.py - agent_utilities/knowledge_graph/standardization/drift.py - agent_utilities/knowledge_graph/standardization/standards.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/models/knowledge_graph.py doc: populated at import with real built-in standards, never empty - id: KG-2.50 name: situate each chunk within the document and embed context+chunk for better recall pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/ingestion/engine.py - agent_utilities/knowledge_graph/ontology/contextual_enrichment.py - agent_utilities/knowledge_graph/ontology/document_processing.py - agent_utilities/mcp/tools/ontology_tools.py doc: situate each chunk within the document and embed context+chunk for better recall - id: KG-2.51 name: a label-scoped MATCH pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/backends/epistemic_graph_backend.py - agent_utilities/knowledge_graph/core/graph_compute.py doc: a label-scoped MATCH - id: KG-2.52 name: the authoritative TBox the platform ships pillar: KG-2 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/ontology_publisher.py doc: the authoritative TBox the platform ships - id: KG-2.53 name: the descriptive process world gains step-level shape pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/enrichment/extractors/aris.py - agent_utilities/knowledge_graph/enrichment/extractors/camunda.py - agent_utilities/knowledge_graph/enrichment/models.py doc: the descriptive process world gains step-level shape - id: KG-2.54 name: Cross-host-safe KG task queue with atomic SKIP LOCKED claims and visibility-timeout recovery on the shared state store pillar: KG-2 status: live code_paths: - agent_utilities/core/leadership.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/postgres_queue_backend.py - agent_utilities/knowledge_graph/ingest_worker.py doc: Cross-host-safe KG task queue with atomic SKIP LOCKED claims and visibility-timeout recovery on the shared state store - id: KG-2.55 name: Fail-loud selectable ingest task-queue backend with explicit kafka and postgres and auto sqlite modes pillar: KG-2 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/__main__.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/kafka_queue_backend.py - agent_utilities/knowledge_graph/core/queue_backend.py - agent_utilities/knowledge_graph/core/shard_topology.py - agent_utilities/knowledge_graph/ingest_worker.py - agent_utilities/orchestration/agent_dispatch.py - agent_utilities/orchestration/scaling_signals.py doc: Fail-loud selectable ingest task-queue backend with explicit kafka and postgres and auto sqlite modes - id: KG-2.56 name: Keyed ingest partitions for per-tenant and per-repo ordering pillar: KG-2 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/core/kafka_queue_backend.py - agent_utilities/knowledge_graph/ingest_worker.py doc: Keyed ingest partitions for per-tenant and per-repo ordering - id: KG-2.57 name: Decoupled kg-ingest consumer group with idempotent at-least-once task claims pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/kafka_queue_backend.py - agent_utilities/knowledge_graph/ingest_worker.py - agent_utilities/knowledge_graph/ingestion/batch_orchestrator.py - agent_utilities/observability/gateway_metrics.py doc: Decoupled kg-ingest consumer group with idempotent at-least-once task claims - id: KG-2.58 name: Tenant-partitioned engine sharding with HRW graph-to-shard routing and tenant to named-graph placement over GRAPH_SERVICE_ENDPOINTS pillar: KG-2 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/__init__.py - agent_utilities/knowledge_graph/core/connection_registry.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/graph_compute.py - agent_utilities/knowledge_graph/core/ingest_engine.py - agent_utilities/knowledge_graph/core/shard_topology.py - agent_utilities/knowledge_graph/distillation/skill_graph_distiller.py - agent_utilities/knowledge_graph/facade.py - agent_utilities/knowledge_graph/ingestion/engine.py - agent_utilities/observability/gateway_metrics.py doc: Tenant-partitioned engine sharding with HRW graph-to-shard routing and tenant to named-graph placement over GRAPH_SERVICE_ENDPOINTS - id: KG-2.59 name: MCP Tool Source Connector pillar: KG-2 status: live code_paths: - agent_utilities/protocols/source_connectors/connectors/__init__.py - agent_utilities/protocols/source_connectors/connectors/database.py - agent_utilities/protocols/source_connectors/connectors/mcp_tool.py doc: MCP Tool Source Connector - id: KG-2.60 name: Data is private-to-its-owner by default; this is the explicit promotion path. action= pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/backends/brain_guarded_backend.py - agent_utilities/knowledge_graph/core/graph_compute.py - agent_utilities/knowledge_graph/core/tenant_sharing.py - agent_utilities/mcp/tools/ontology_tools.py doc: Data is private-to-its-owner by default; this is the explicit promotion path. action= - id: KG-2.61 name: ── pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/backends/postgresql_backend.py doc: ── - id: KG-2.62 name: The elastic layer over KG-2.58 shard routing pillar: KG-2 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/core/tenant_engine_pool.py doc: The elastic layer over KG-2.58 shard routing - id: KG-2.63 name: Profile / imprint / ontology-map a registered external graph pillar: KG-2 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/deployment/doctor.py - agent_utilities/knowledge_graph/backends/__init__.py - agent_utilities/knowledge_graph/backends/age_backend.py - agent_utilities/knowledge_graph/backends/base.py - agent_utilities/knowledge_graph/backends/epistemic_graph_backend.py - agent_utilities/knowledge_graph/backends/fanout_backend.py - agent_utilities/knowledge_graph/backends/postgresql_backend.py - agent_utilities/knowledge_graph/backends/sparql/stardog_backend.py - agent_utilities/knowledge_graph/core/connection_profiler.py - agent_utilities/knowledge_graph/core/connection_registry.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/mcp/tools/query_tools.py - agent_utilities/mcp/tools/write_ingest_tools.py doc: Profile / imprint / ontology-map a registered external graph - id: KG-2.64 name: document → atomic-triple fact extraction pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/extraction/__init__.py - agent_utilities/knowledge_graph/extraction/fact_extractor.py - agent_utilities/mcp/tools/write_ingest_tools.py - agent_utilities/runtime/provenance.py doc: document → atomic-triple fact extraction - id: KG-2.65 name: GPU-slot-scheduled fact extraction. Unlike the pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/ingestion/gpu_slot_scheduler.py - agent_utilities/mcp/tools/write_ingest_tools.py - agent_utilities/tools/code_intelligence_tools.py doc: GPU-slot-scheduled fact extraction. Unlike the - id: KG-2.66 name: Keys are normalised lowercase; an extension is stored with a pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/extraction/readers.py - agent_utilities/knowledge_graph/extraction/readers_office.py - agent_utilities/knowledge_graph/ingestion/engine.py - agent_utilities/protocols/source_connectors/connectors/__init__.py - agent_utilities/protocols/source_connectors/connectors/reader.py doc: Keys are normalised lowercase; an extension is stored with a - id: KG-2.67 name: a first-class action-conditioned world model wrapping the graph pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/world_model.py doc: a first-class action-conditioned world model wrapping the graph - id: KG-2.68 name: a first-class Reasoner paradigm abstraction with an outcome-learning router that selects a reasoning paradigm per task by capability tags blended with a learned reward EMA and feeds the scored result back so routing self-improves pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/reasoner.py doc: a first-class Reasoner paradigm abstraction with an outcome-learning router that selects a reasoning paradigm per task by capability tags blended with a learned reward EMA and feeds the scored result back so routing self-improves - id: KG-2.69 name: inductive program-synthesis search over a typed primitive composition DSL biased by a minimum-description-length prior so among programs that fit the examples the shortest is preferred a finite stand-in for the Solomonoff universal prior pillar: KG-2 status: live code_paths: - agent_utilities/harness/program_synthesis.py - agent_utilities/harness/selection_operators.py doc: inductive program-synthesis search over a typed primitive composition DSL biased by a minimum-description-length prior so among programs that fit the examples the shortest is preferred a finite stand-in for the Solomonoff universal prior - id: KG-2.70 name: /2.71/2.72, AHE-3.30; distills arXiv pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/research/loop_controller.py - agent_utilities/knowledge_graph/search_synthesis/__init__.py - agent_utilities/knowledge_graph/search_synthesis/evidence_subgraph.py - agent_utilities/knowledge_graph/search_synthesis/models.py - agent_utilities/mcp/tools/query_tools.py doc: /2.71/2.72, AHE-3.30; distills arXiv - id: KG-2.71 name: and formulate + adversarially refine a pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/search_synthesis/__init__.py - agent_utilities/knowledge_graph/search_synthesis/shortcut_risks.py doc: and formulate + adversarially refine a - id: KG-2.72 name: question formulation + adversarial refinement pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/search_synthesis/__init__.py - agent_utilities/knowledge_graph/search_synthesis/question_formulation.py doc: question formulation + adversarial refinement - id: KG-2.73 name: KG-2.73 pillar: KG-2 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/world_model_task.py - agent_utilities/knowledge_graph/core/world_model.py doc: KG-2.73 - id: KG-2.73b name: forward-simulate the learned world model with pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/world_model.py - agent_utilities/mcp/tools/analysis_tools.py doc: forward-simulate the learned world model with - id: KG-2.74 name: Concurrent N-way mirroring health/repair ── pillar: KG-2 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/backends/__init__.py - agent_utilities/knowledge_graph/backends/base.py - agent_utilities/knowledge_graph/backends/contrib/ladybug_backend.py - agent_utilities/knowledge_graph/backends/fanout_backend.py - agent_utilities/knowledge_graph/backends/outbox.py - agent_utilities/knowledge_graph/backends/sparql/stardog_backend.py - agent_utilities/knowledge_graph/extraction/ontology_grounding.py - agent_utilities/knowledge_graph/extraction/readers_media.py - agent_utilities/knowledge_graph/extraction/structure_router.py - agent_utilities/knowledge_graph/migration.py - agent_utilities/mcp/tools/analysis_tools.py doc: Concurrent N-way mirroring health/repair ── - id: KG-2.75 name: KG-2.75 pillar: KG-2 status: live code_paths: - agent_utilities/automation/research_pipeline.py - agent_utilities/knowledge_graph/assimilation/concept_matcher.py - agent_utilities/knowledge_graph/assimilation/ingest.py - agent_utilities/knowledge_graph/research/loop_controller.py - agent_utilities/models/knowledge_graph.py doc: KG-2.75 - id: KG-2.76 name: KG-2.76 pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/research_objects.py doc: KG-2.76 - id: KG-2.77 name: KG-2.77 pillar: KG-2 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/research/loop_controller.py - agent_utilities/knowledge_graph/research/research_intelligence_engine.py doc: KG-2.77 - id: KG-2.78 name: These properties carry the pillar: KG-2 status: live code_paths: - agent_utilities/core/sessions.py - agent_utilities/knowledge_graph/research/loop_controller.py - agent_utilities/knowledge_graph/research/loops.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/mcp/tools/state_tools.py - agent_utilities/orchestration/agent_dispatch_worker.py doc: These properties carry the - id: KG-2.79 name: Best-effort + lightweight; never blocks pillar: KG-2 status: live code_paths: - agent_utilities/gateway/graph_api.py - agent_utilities/knowledge_graph/research/ara/__init__.py - agent_utilities/knowledge_graph/research/ara/reasoning_driver.py - agent_utilities/knowledge_graph/research/loop_controller.py - agent_utilities/mcp/tools/state_tools.py doc: Best-effort + lightweight; never blocks - id: KG-2.80 name: verified_by/implemented_by/contains/ pillar: KG-2 status: live code_paths: - agent_utilities/gateway/research_api.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/ontology/interfaces.py - agent_utilities/knowledge_graph/ontology/links.py - agent_utilities/knowledge_graph/ontology/value_types.py - agent_utilities/knowledge_graph/research/ara/artifact.py - agent_utilities/knowledge_graph/research/ara/compiler.py - agent_utilities/knowledge_graph/research/ara/exploration.py - agent_utilities/knowledge_graph/research/ara/live_manager.py - agent_utilities/knowledge_graph/research/ara/seal.py - agent_utilities/knowledge_graph/research/ara/service.py - agent_utilities/models/knowledge_graph.py doc: verified_by/implemented_by/contains/ - id: KG-2.81 name: Kyle insider/stealth surveillance signal ontology pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/ontology/__init__.py - agent_utilities/knowledge_graph/ontology/finance_objects.py doc: Kyle insider/stealth surveillance signal ontology - id: KG-2.82 name: / AHE-3.33 — record this base pillar: KG-2 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/agentic_evolution_engine.py - agent_utilities/harness/decentralized_memory.py doc: / AHE-3.33 — record this base - id: KG-2.83 name: explicit node↔node contradiction/friction surface pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/adaptation/contradiction_detector.py - agent_utilities/knowledge_graph/research/night_shift.py - agent_utilities/mcp/tools/analysis_tools.py doc: explicit node↔node contradiction/friction surface - id: KG-2.84 name: run one autonomous night-shift cycle over a pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/research/night_shift.py - agent_utilities/mcp/tools/analysis_tools.py doc: run one autonomous night-shift cycle over a - id: KG-2.85 name: Adaptive Chunk Selection pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/knowledge_graph/retrieval/neural_reranker.py - agent_utilities/knowledge_graph/retrieval/reasoning_reranker.py - agent_utilities/knowledge_graph/retrieval/score_gate.py doc: Adaptive Chunk Selection - id: KG-2.86 name: ChronoID. Fits a residual-quantization codebook on the pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/knowledge_graph/retrieval/generative_recommender.py - agent_utilities/knowledge_graph/retrieval/temporal_semantic_id.py - agent_utilities/mcp/tools/query_tools.py doc: ChronoID. Fits a residual-quantization codebook on the - id: KG-2.87 name: Adaptive Stopping for Iterative Retrieval pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/knowledge_graph/retrieval/adaptive_stopping.py doc: Adaptive Stopping for Iterative Retrieval - id: KG-2.88 name: assimilated from ADORE pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/orchestration/engine_query.py - agent_utilities/knowledge_graph/retrieval/iterative_expansion.py - agent_utilities/mcp/tools/query_tools.py doc: assimilated from ADORE - id: KG-2.89 name: Role-aware multi-database registry plus live config mutation pillar: KG-2 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/deployment/config_generator.py - agent_utilities/knowledge_graph/backends/__init__.py - agent_utilities/knowledge_graph/core/connection_registry.py - agent_utilities/knowledge_graph/setup/database_environment.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/mcp/tools/write_ingest_tools.py doc: Role-aware multi-database registry plus live config mutation - id: KG-2.90 name: / KG-2.91 — a KG-native, propose-only background distiller that pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/distillation/skill_synthesizer.py - agent_utilities/knowledge_graph/ontology/interfaces.py - agent_utilities/knowledge_graph/research/loop_controller.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/models/knowledge_graph.py doc: / KG-2.91 — a KG-native, propose-only background distiller that - id: KG-2.91 name: AUTOMATES pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/models/knowledge_graph.py doc: AUTOMATES - id: KG-2.92 name: Monte-Carlo GRAPH search code evolution pillar: KG-2 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/agentic_evolution_engine.py - agent_utilities/harness/graph_search_evolution.py - agent_utilities/mcp/tools/analysis_tools.py doc: Monte-Carlo GRAPH search code evolution - id: KG-2.93 name: PauseRec implicit-reasoning generative recommender pillar: KG-2 status: live code_paths: - agent_utilities/harness/assimilation_benchmark.py - agent_utilities/knowledge_graph/retrieval/generative_recommender.py - agent_utilities/knowledge_graph/retrieval/pause_token_trainer.py - agent_utilities/mcp/tools/analysis_tools.py doc: PauseRec implicit-reasoning generative recommender - id: KG-2.94 name: LLM inference sampling knobs temperature top_p top_k min_p repetition_penalty max_tokens and penalties as SHACL-bounded ontology value types plus the two-surface sampling-profile tool pillar: KG-2 status: live code_paths: - agent_utilities/agent/sampling_profile.py - agent_utilities/knowledge_graph/ontology/value_types.py doc: LLM inference sampling knobs temperature top_p top_k min_p repetition_penalty max_tokens and penalties as SHACL-bounded ontology value types plus the two-surface sampling-profile tool - id: KG-2.95 name: InferenceProfile + SamplingConfigurable ontology interfaces and a first-class Model object type projecting configured models and their tuned sampling profiles to OWL pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/ontology/interfaces.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/model_registry.py doc: InferenceProfile + SamplingConfigurable ontology interfaces and a first-class Model object type projecting configured models and their tuned sampling profiles to OWL - id: KG-2.96 name: Typed ontology links binding Model/InferenceProfile to TaskClass/Role/Agent pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/ontology/links.py - agent_utilities/models/knowledge_graph.py doc: Typed ontology links binding Model/InferenceProfile to TaskClass/Role/Agent - id: KG-2.97 name: ingest the universal-skills workflow corpus as pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/ingestion/skill_workflow_ingest.py - agent_utilities/mcp/tools/write_ingest_tools.py doc: ingest the universal-skills workflow corpus as - id: KG-2.98 name: One source→ pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/etl/__init__.py - agent_utilities/knowledge_graph/etl/pipeline.py - agent_utilities/mcp/tools/ontology_tools.py doc: One source→ - id: KG-2.99 name: KG-2.99 pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/etl/lineage.py - agent_utilities/mcp/tools/ontology_tools.py doc: KG-2.99 - id: KG-2.100 name: the type/scope-resolved call/inheritance graph pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/enrichment/extractors/code_test.py - agent_utilities/knowledge_graph/enrichment/pipeline.py - agent_utilities/knowledge_graph/ingestion/engine.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/analysis_tools.py doc: the type/scope-resolved call/inheritance graph - id: KG-2.101 name: model-free similar-code lookup. Returns the pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/analysis_tools.py doc: model-free similar-code lookup. Returns the - id: KG-2.102 name: HTTP routes from handler decorators pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/enrichment/extractors/code_test.py - agent_utilities/knowledge_graph/enrichment/pipeline.py - agent_utilities/knowledge_graph/enrichment/routes.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/analysis_tools.py doc: HTTP routes from handler decorators - id: KG-2.103 name: read them here so the pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/enrichment/iac.py - agent_utilities/knowledge_graph/enrichment/pipeline.py doc: read them here so the - id: KG-2.104 name: mine git history for files that change together pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/enrichment/git_coupling.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/analysis_tools.py doc: mine git history for files that change together - id: KG-2.105 name: Architecture Decision Record CRUD pillar: KG-2 status: live code_paths: - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/analysis_tools.py doc: Architecture Decision Record CRUD - id: KG-2.106 name: engine built with ast-extended pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/gitlab_indexer.py - agent_utilities/knowledge_graph/enrichment/pipeline.py doc: engine built with ast-extended - id: KG-2.107 name: KG-2.107 pillar: KG-2 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py doc: KG-2.107 - id: KG-2.108 name: KG-2.108 pillar: KG-2 status: live code_paths: - agent_utilities/harness/harness_grounding.py - agent_utilities/protocols/source_connectors/connectors/mcp_tool.py doc: KG-2.108 - id: KG-2.109 name: KG-2.109 pillar: KG-2 status: live code_paths: - agent_utilities/harness/harness_gate.py - agent_utilities/knowledge_graph/core/owl_bridge.py doc: KG-2.109 - id: LGC-1.0 name: LGC-1.0 pillar: LGC-1 status: live code_paths: - agent_utilities/knowledge_graph/backends/owl/owlready2_backend.py - agent_utilities/knowledge_graph/core/owl_bridge.py doc: LGC-1.0 - id: ML-011 name: join inference pillar: ML-011 status: live code_paths: - agent_utilities/knowledge_graph/ingestion/gpu_slot_scheduler.py doc: join inference - id: ORCH-1.0 name: Inject signal board observations from prior adaptive_agent_router pillar: ORCH-1 status: live code_paths: - agent_utilities/__main__.py - agent_utilities/capabilities/adversarial_verifier.py - agent_utilities/graph/_router_impl.py - agent_utilities/graph/adaptive_agent_router.py - agent_utilities/graph/executor.py - agent_utilities/graph/hierarchical_planner.py - agent_utilities/graph/hsm.py - agent_utilities/graph/lifecycle.py - agent_utilities/graph/models.py - agent_utilities/graph/state.py - agent_utilities/harness/variant_pool.py - agent_utilities/models/knowledge_graph.py - agent_utilities/orchestration/engine.py - agent_utilities/orchestration/protocol.py - agent_utilities/prompting/structured.py - agent_utilities/protocols/agui_emitter.py doc: Inject signal board observations from prior adaptive_agent_router - id: ORCH-1.1 name: Current nesting depth for recursive graph orchestration pillar: ORCH-1 status: live code_paths: - agent_utilities/core/checkpoint/manager.py - agent_utilities/core/registry/kg_adapter.py - agent_utilities/graph/_router_impl.py - agent_utilities/graph/executor.py - agent_utilities/graph/hierarchical_planner.py - agent_utilities/graph/horizon_curriculum.py - agent_utilities/graph/nodes.py - agent_utilities/graph/routing/strategies/llm_planner.py - agent_utilities/graph/state.py - agent_utilities/graph/subagent_patterns.py - agent_utilities/graph/team_composer.py - agent_utilities/graph/verification.py - agent_utilities/harness/continuous_evaluation_engine.py - agent_utilities/harness/evolve_agent.py - agent_utilities/harness/variant_pool.py - agent_utilities/models/graph.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/sdd.py - agent_utilities/patterns/__init__.py - agent_utilities/patterns/prioritization.py - agent_utilities/patterns/prompt_chain.py - agent_utilities/rlm/__init__.py - agent_utilities/rlm/config.py - agent_utilities/rlm/hook.py - agent_utilities/rlm/repl.py - agent_utilities/tools/db_tools.py doc: Current nesting depth for recursive graph orchestration - id: ORCH-1.2 name: Invalidate hot cache so routing reflects new self-knowledge pillar: ORCH-1 status: live code_paths: - agent_utilities/agent/discovery.py - agent_utilities/agent/registry_builder.py - agent_utilities/capabilities/adversarial_verifier.py - agent_utilities/core/config.py - agent_utilities/core/default_catalog.py - agent_utilities/core/model_factory.py - agent_utilities/core/registry/kg_adapter.py - agent_utilities/core/registry/package_adapter.py - agent_utilities/core/resource_optimizer.py - agent_utilities/graph/_router_impl.py - agent_utilities/graph/adaptive_agent_router.py - agent_utilities/graph/executor.py - agent_utilities/graph/hierarchical_planner.py - agent_utilities/graph/parallel_engine.py - agent_utilities/graph/routing/strategies/learned.py - agent_utilities/graph/state.py - agent_utilities/graph/topology_engine.py - agent_utilities/graph/workspace_attention.py - agent_utilities/knowledge_graph/core/engine_memory.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/knowledge_graph/pipeline/runner.py - agent_utilities/knowledge_graph/retrieval/memory_retriever.py - agent_utilities/mcp/agent_manager.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/mcp.py - agent_utilities/models/model_registry.py - agent_utilities/protocols/a2a_config.py - agent_utilities/workflows/skill_picker.py doc: Invalidate hot cache so routing reflects new self-knowledge - id: ORCH-1.3 name: Visibility allow-list of upstream step ids whose results pillar: ORCH-1 status: live code_paths: - agent_utilities/base_utilities.py - agent_utilities/capabilities/checkpointing.py - agent_utilities/core/checkpoint/manager.py - agent_utilities/graph/_router_impl.py - agent_utilities/graph/coordination.py - agent_utilities/graph/executor.py - agent_utilities/graph/hierarchical_planner.py - agent_utilities/graph/state.py - agent_utilities/graph/subagent_patterns.py - agent_utilities/graph/workspace_attention.py - agent_utilities/http/__init__.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/sdd.py - agent_utilities/models/usage.py - agent_utilities/orchestration/engine.py - agent_utilities/protocols/a2a_graph_skill.py - agent_utilities/sdd/orchestrator.py - agent_utilities/security/__init__.py - agent_utilities/security/execution_stability_engine.py - agent_utilities/tools/sdd_tools.py doc: Visibility allow-list of upstream step ids whose results - id: ORCH-1.3b name: ORCH-1.3b pillar: ORCH-1 status: live code_paths: - agent_utilities/graph/subagent_patterns.py doc: ORCH-1.3b - id: ORCH-1.4 name: Session ID of the parent graph if this state was forked pillar: ORCH-1 status: live code_paths: - agent_utilities/capabilities/hooks.py - agent_utilities/core/config.py - agent_utilities/core/registry/kg_adapter.py - agent_utilities/core/registry/service_adapter.py - agent_utilities/graph/__init__.py - agent_utilities/graph/_router_impl.py - agent_utilities/graph/adapters/external_plugin_adapter.py - agent_utilities/graph/adaptive_agent_router.py - agent_utilities/graph/horizon_curriculum.py - agent_utilities/graph/kg_graph_factory.py - agent_utilities/graph/persistent_agents.py - agent_utilities/graph/protocol_agnostic_execution.py - agent_utilities/graph/routing/strategies/policy.py - agent_utilities/graph/state.py - agent_utilities/graph_orchestration.py - agent_utilities/knowledge_graph/core/engine.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/schema_definition.py doc: Session ID of the parent graph if this state was forked - id: ORCH-1.8 name: Dependency cycle detected — falling back pillar: ORCH-1 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/core/execution/protocol.py - agent_utilities/graph/__init__.py - agent_utilities/graph/gather.py - agent_utilities/graph/manifest_generators.py - agent_utilities/graph/parallel_engine.py - agent_utilities/graph/verification.py - agent_utilities/harness/evolve_agent.py - agent_utilities/knowledge_graph/dspy_kg_bridge.py - agent_utilities/models/execution_manifest.py - agent_utilities/orchestration/engine.py - agent_utilities/workflows/distillation_hook.py - agent_utilities/workflows/runner.py - agent_utilities/workflows/skill_compiler.py - agent_utilities/workflows/visualizer.py doc: Dependency cycle detected — falling back - id: ORCH-1.9 name: Autonomous Department Orchestration pillar: ORCH-1 status: live code_paths: - agent_utilities/graph/manifest_generators.py - agent_utilities/models/company.py doc: Autonomous Department Orchestration - id: ORCH-1.10 name: Graph-Native Reactive Event Sourcing and OS Guardrails pillar: ORCH-1 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/graph/__init__.py - agent_utilities/graph/reactive/__init__.py - agent_utilities/graph/reactive/dispatcher.py - agent_utilities/graph/reactive/ledger.py - agent_utilities/knowledge_graph/core/event_backend.py - agent_utilities/knowledge_graph/core/kafka_queue_backend.py - agent_utilities/knowledge_graph/core/queue_backend.py doc: Graph-Native Reactive Event Sourcing and OS Guardrails - id: ORCH-1.11 name: Compiled Orchestration Kernel pillar: ORCH-1 status: live code_paths: - agent_utilities/core/__init__.py - agent_utilities/core/wasm_runner.py - agent_utilities/knowledge_graph/core/graph_compute.py doc: Compiled Orchestration Kernel - id: ORCH-1.12 name: run the Predict-RLM runtime on an ad-hoc task pillar: ORCH-1 status: live code_paths: - agent_utilities/graph/_router_impl.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/rlm/config.py - agent_utilities/rlm/predict_rlm.py - agent_utilities/rlm/repl.py - agent_utilities/rlm/runner.py - agent_utilities/rlm/schema.py doc: run the Predict-RLM runtime on an ad-hoc task - id: ORCH-1.13 name: optimize a skill prompt via the GEPA loop pillar: ORCH-1 status: live code_paths: - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/models/knowledge_graph.py - agent_utilities/rlm/gepa.py - agent_utilities/rlm/runner.py doc: optimize a skill prompt via the GEPA loop - id: ORCH-1.20 name: Service Registry Initialization pillar: ORCH-1 status: live code_paths: - agent_utilities/orchestration/engine.py doc: Service Registry Initialization - id: ORCH-1.21 name: when a remote MCP child fails, anyio wraps the real cause pillar: ORCH-1 status: live code_paths: - agent_utilities/orchestration/agent_runner.py doc: when a remote MCP child fails, anyio wraps the real cause - id: ORCH-1.22 name: Workflow Persistence & Replay Relationships pillar: ORCH-1 status: live code_paths: - agent_utilities/knowledge_graph/workflow_store.py - agent_utilities/models/schema_definition.py doc: Workflow Persistence & Replay Relationships - id: ORCH-1.23 name: Semantic Workflow Retrieval pillar: ORCH-1 status: live code_paths: - agent_utilities/knowledge_graph/workflow_compiler.py doc: Semantic Workflow Retrieval - id: ORCH-1.24 name: Workflow Lifecycle Management pillar: ORCH-1 status: live code_paths: - agent_utilities/knowledge_graph/kb/x_workflows.py - agent_utilities/orchestration/engine.py - agent_utilities/workflows/__init__.py - agent_utilities/workflows/runner.py doc: Workflow Lifecycle Management - id: ORCH-1.26 name: RLM synthesis failed, falling back to pillar: ORCH-1 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/graph/parallel_engine.py - agent_utilities/models/execution_manifest.py doc: RLM synthesis failed, falling back to - id: ORCH-1.27 name: a Conductor-assigned per-step model_id wins over both the pillar: ORCH-1 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/core/model_factory.py - agent_utilities/graph/executor.py - agent_utilities/graph/routing/strategies/llm_planner.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/models/model_registry.py - agent_utilities/models/sdd.py - agent_utilities/rlm/gepa.py - agent_utilities/rlm/roles.py doc: a Conductor-assigned per-step model_id wins over both the - id: ORCH-1.28 name: Composable Skills + Generic Environment Adapter pillar: ORCH-1 status: live code_paths: - agent_utilities/rlm/predict_rlm.py - agent_utilities/rlm/runner.py - agent_utilities/rlm/skills.py doc: Composable Skills + Generic Environment Adapter - id: ORCH-1.29 name: populate a structured RunTrace as the live loop runs, so the GEPA pillar: ORCH-1 status: live code_paths: - agent_utilities/rlm/repl.py - agent_utilities/rlm/runner.py - agent_utilities/rlm/telemetry.py doc: populate a structured RunTrace as the live loop runs, so the GEPA - id: ORCH-1.30 name: selection happens on UNSEEN data, so a candidate that merely memorized the pillar: ORCH-1 status: live code_paths: - agent_utilities/graph/coordination.py - agent_utilities/harness/__init__.py - agent_utilities/harness/selection_operators.py - agent_utilities/harness/variant_pool.py - agent_utilities/rlm/dynamic_reward.py - agent_utilities/rlm/gepa.py doc: selection happens on UNSEEN data, so a candidate that merely memorized the - id: ORCH-1.31 name: Graph-Native Optimization State ────────────────────────── pillar: ORCH-1 status: live code_paths: - agent_utilities/rlm/gepa.py doc: Graph-Native Optimization State ────────────────────────── - id: ORCH-1.32 name: model the wave as a Multi-Agent Social System and pillar: ORCH-1 status: live code_paths: - agent_utilities/graph/parallel_engine.py - agent_utilities/graph/social_system.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/models/execution_manifest.py doc: model the wave as a Multi-Agent Social System and - id: ORCH-1.33 name: additive multi-CLI adapter dispatch. When a manifest requests an external runtime pillar: ORCH-1 status: live code_paths: - agent_utilities/core/execution/adapters/__init__.py - agent_utilities/core/execution/adapters/base.py - agent_utilities/core/execution/adapters/defs/__init__.py - agent_utilities/core/execution/adapters/executor.py - agent_utilities/core/execution/adapters/registry.py - agent_utilities/core/execution/engine.py - agent_utilities/core/execution/stream_handlers.py doc: additive multi-CLI adapter dispatch. When a manifest requests an external runtime - id: ORCH-1.34 name: BYOK custom endpoint. The provider proxy emits OpenAI-compatible pillar: ORCH-1 status: live code_paths: - agent_utilities/core/credentials.py - agent_utilities/core/execution/provider_proxy.py - agent_utilities/core/model_factory.py - agent_utilities/security/egress.py - agent_utilities/server/app.py - agent_utilities/server/routers/proxy.py doc: BYOK custom endpoint. The provider proxy emits OpenAI-compatible - id: ORCH-1.35 name: Held-turn registry for mid-turn tool-result injection pillar: ORCH-1 status: live code_paths: - agent_utilities/core/execution/held_turns.py - agent_utilities/security/sidecar_runtime.py - agent_utilities/server/routers/human.py doc: Held-turn registry for mid-turn tool-result injection - id: ORCH-1.36 name: retry_on predicate raised; treating as non-retryable pillar: ORCH-1 status: live code_paths: - agent_utilities/core/http_client.py - agent_utilities/graph/executor.py - agent_utilities/graph/parallel_engine.py - agent_utilities/http/__init__.py - agent_utilities/knowledge_graph/backends/contrib/ladybug_backend.py - agent_utilities/knowledge_graph/backends/postgresql_backend.py - agent_utilities/knowledge_graph/memory/learning_engine.py - agent_utilities/orchestration/resilience.py - agent_utilities/patterns/prompt_chain.py doc: retry_on predicate raised; treating as non-retryable - id: ORCH-1.37 name: Orchestration execution-flow mermaid-diagram surfacing in graph_orchestrate responses pillar: ORCH-1 status: live code_paths: - agent_utilities/graph/_router_impl.py - agent_utilities/graph/executor.py - agent_utilities/graph/hierarchical_planner.py - agent_utilities/graph/verification.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/orchestration/agent_runner.py - agent_utilities/orchestration/manager.py doc: Orchestration execution-flow mermaid-diagram surfacing in graph_orchestrate responses - id: ORCH-1.38 name: Tiered RLM code sandbox with a uniform Sandbox contract and a capability router that escalates across local, monty, wasm, and docker backends pillar: ORCH-1 status: live code_paths: - agent_utilities/knowledge_graph/research/change_synthesis.py - agent_utilities/rlm/config.py - agent_utilities/rlm/repl.py - agent_utilities/rlm/sandboxes/__init__.py - agent_utilities/rlm/sandboxes/analyzer.py - agent_utilities/rlm/sandboxes/base.py - agent_utilities/rlm/sandboxes/docker_backend.py - agent_utilities/rlm/sandboxes/local_backend.py - agent_utilities/rlm/sandboxes/monty_backend.py - agent_utilities/rlm/sandboxes/registry.py - agent_utilities/rlm/sandboxes/router.py - agent_utilities/rlm/sandboxes/wasm_backend.py - agent_utilities/rlm/telemetry.py doc: Tiered RLM code sandbox with a uniform Sandbox contract and a capability router that escalates across local, monty, wasm, and docker backends - id: ORCH-1.39 name: Invoker→spawned-agent handoff of curated context, token budget, tool scope and credential reference pillar: ORCH-1 status: live code_paths: - agent_utilities/graph/_router_impl.py - agent_utilities/graph/executor.py - agent_utilities/graph/state.py - agent_utilities/knowledge_graph/core/engine.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/mcp/tools/query_tools.py - agent_utilities/orchestration/agent_runner.py - agent_utilities/orchestration/engine.py - agent_utilities/orchestration/manager.py doc: Invoker→spawned-agent handoff of curated context, token budget, tool scope and credential reference - id: ORCH-1.40 name: Session-anchored collections and native invoker↔spawned-agent message channels pillar: ORCH-1 status: live code_paths: - agent_utilities/graph/executor.py - agent_utilities/graph/state.py - agent_utilities/knowledge_graph/backends/epistemic_graph_backend.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/mcp/tools/query_tools.py - agent_utilities/messaging/agent_channel.py - agent_utilities/models/agent.py - agent_utilities/orchestration/agent_runner.py - agent_utilities/orchestration/engine.py doc: Session-anchored collections and native invoker↔spawned-agent message channels - id: ORCH-1.41 name: REST twin of graph_orchestrate compile_process pillar: ORCH-1 status: live code_paths: - agent_utilities/knowledge_graph/process_plan_compiler.py - agent_utilities/mcp/kg_server.py - agent_utilities/mcp/tools/analysis_tools.py doc: REST twin of graph_orchestrate compile_process - id: ORCH-1.42 name: Ontology Validation on the Execution Path pillar: ORCH-1 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/core/workflow_gate.py - agent_utilities/mcp/tools/analysis_tools.py doc: Ontology Validation on the Execution Path - id: ORCH-1.43 name: close the descriptive↔executable provenance loop pillar: ORCH-1 status: live code_paths: - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/workflows/runner.py doc: close the descriptive↔executable provenance loop - id: ORCH-1.44 name: Durable goal registry — goals persist across restarts and stranded runs rehydrate as orphaned instead of silently vanishing pillar: ORCH-1 status: live code_paths: - agent_utilities/core/sessions.py - agent_utilities/models/goal.py doc: Durable goal registry — goals persist across restarts and stranded runs rehydrate as orphaned instead of silently vanishing - id: ORCH-1.45 name: Queue-driven agent dispatch with session-keyed partitions pillar: ORCH-1 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/core/sessions.py - agent_utilities/gateway/fleet.py - agent_utilities/knowledge_graph/core/kafka_queue_backend.py - agent_utilities/knowledge_graph/core/postgres_queue_backend.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/observability/gateway_metrics.py - agent_utilities/orchestration/agent_dispatch.py - agent_utilities/orchestration/agent_dispatch_worker.py doc: Queue-driven agent dispatch with session-keyed partitions - id: ORCH-1.46 name: Action/Observation event protocol for the developer-workspace runtime pillar: ORCH-1 status: live code_paths: - agent_utilities/runtime/bridge.py - agent_utilities/runtime/events.py doc: Action/Observation event protocol for the developer-workspace runtime - id: ORCH-1.47 name: System prompt for the KG-grounded SWE agent pillar: ORCH-1 status: live code_paths: - agent_utilities/agent/swe_prompts.py - agent_utilities/models/agent.py - agent_utilities/orchestration/engine.py - agent_utilities/orchestration/swe_agent.py - agent_utilities/tools/swe_workspace_tools.py - agent_utilities/tools/tool_registry.py doc: System prompt for the KG-grounded SWE agent - id: ORCH-1.48 name: Token-budgeted, importance-ranked repo-map skeleton for compact context injection pillar: ORCH-1 status: live code_paths: - agent_utilities/models/codemap.py - agent_utilities/server/models.py doc: Token-budgeted, importance-ranked repo-map skeleton for compact context injection - id: ORCH-1.49 name: Robust multi-format edit-application engine with graduated fuzzy matching and a reflection loop pillar: ORCH-1 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/edit_engine.py doc: Robust multi-format edit-application engine with graduated fuzzy matching and a reflection loop - id: ORCH-1.50 name: Task-management ergonomics on SDD with complexity scoring, dependency-aware next-task, scope and tags pillar: ORCH-1 status: live code_paths: - agent_utilities/mcp/harness_server.py - agent_utilities/models/sdd.py - agent_utilities/sdd/__init__.py doc: Task-management ergonomics on SDD with complexity scoring, dependency-aware next-task, scope and tags - id: ORCH-1.51 name: a virtual-agent-economy task allocator that runs a capability-gated second-price auction so a collective self-organizes who does what via price signals instead of static role routing, surfacing scarcity and value pillar: ORCH-1 status: live code_paths: - agent_utilities/orchestration/agent_market.py doc: a virtual-agent-economy task allocator that runs a capability-gated second-price auction so a collective self-organizes who does what via price signals instead of static role routing, surfacing scarcity and value - id: ORCH-1.52 name: an emergent-specialization discovery pass that clusters the failing or expensive task stream and proposes a new specialist archetype for any niche no existing archetype covers, so the collective autonomously increases its division of labour pillar: ORCH-1 status: live code_paths: - agent_utilities/graph/specialization_discovery.py doc: an emergent-specialization discovery pass that clusters the failing or expensive task stream and proposes a new specialist archetype for any niche no existing archetype covers, so the collective autonomously increases its division of labour - id: ORCH-1.53 name: a hierarchical federated coordination protocol that runs consensus within agent neighborhoods and rolls up only cluster representatives, bounding interaction density so a collective coordinates at thousands of agents instead of O pillar: ORCH-1 status: live code_paths: - agent_utilities/graph/hierarchical_coordination.py doc: a hierarchical federated coordination protocol that runs consensus within agent neighborhoods and rolls up only cluster representatives, bounding interaction density so a collective coordinates at thousands of agents instead of O - id: ORCH-1.54 name: Drop-in RLM completion client and model-family-aware REPL prompt pillar: ORCH-1 status: live code_paths: - agent_utilities/harness/agentic_evolution_engine.py - agent_utilities/mcp/tools/analysis_tools.py - agent_utilities/rlm/__init__.py - agent_utilities/rlm/client.py - agent_utilities/rlm/config.py - agent_utilities/rlm/prompts.py - agent_utilities/rlm/repl.py doc: Drop-in RLM completion client and model-family-aware REPL prompt - id: ORCH-1.55 name: Eval-Set Optimization & Compounding Learning Loop pillar: ORCH-1 status: live code_paths: - agent_utilities/harness/continuous_evaluation_engine.py - agent_utilities/rlm/eval_set_optimizer.py doc: Eval-Set Optimization & Compounding Learning Loop - id: ORCH-1.56 name: feed the cycle outcome to the Fast-Slow controller pillar: ORCH-1 status: live code_paths: - agent_utilities/harness/agentic_evolution_engine.py - agent_utilities/harness/fast_slow_controller.py doc: feed the cycle outcome to the Fast-Slow controller - id: ORCH-1.57 name: Substrate training-job emission pillar: ORCH-1 status: live code_paths: - agent_utilities/harness/agentic_evolution_engine.py - agent_utilities/harness/substrate_trainer.py doc: Substrate training-job emission - id: ORCH-1.58 name: Task-aware per-call LLM sampling that selects and threads a SamplingProfile of temperature top_p top_k min_p repetition_penalty max_tokens and penalties per question by task-class and role pillar: ORCH-1 status: live code_paths: - agent_utilities/agent/factory.py - agent_utilities/agent/sampling_profile.py - agent_utilities/gateway/ontology_api.py - agent_utilities/graph/adaptive_agent_router.py - agent_utilities/mcp/tools/ontology_tools.py - agent_utilities/models/model_registry.py - agent_utilities/rlm/repl.py doc: Task-aware per-call LLM sampling that selects and threads a SamplingProfile of temperature top_p top_k min_p repetition_penalty max_tokens and penalties per question by task-class and role - id: ORCH-2.0 name: Orchestration Engine pillar: ORCH-2 status: live code_paths: - agent_utilities/orchestration/__init__.py - agent_utilities/orchestration/engine.py doc: Orchestration Engine - id: ORCH-5.0 name: Durable session and autonomous goal persistence with iterative background goal loops pillar: ORCH-5 status: live code_paths: - agent_utilities/core/sessions.py doc: Durable session and autonomous goal persistence with iterative background goal loops - id: OS-5.0 name: FileWatcher — watchdog-triggered graph execution pillar: OS-5 status: live code_paths: - agent_utilities/agent/factory.py - agent_utilities/automation/__init__.py - agent_utilities/automation/file_watcher.py - agent_utilities/automation/maintenance_cron.py - agent_utilities/core/config.py - agent_utilities/core/discovery.py - agent_utilities/core/paths.py - agent_utilities/core/registry/package_adapter.py - agent_utilities/core/workspace.py - agent_utilities/graph/_router_impl.py - agent_utilities/graph/state.py - agent_utilities/knowledge_graph/backends/epistemic_graph_backend.py - agent_utilities/knowledge_graph/backends/postgresql_backend.py - agent_utilities/models/knowledge_graph.py - agent_utilities/security/execution_stability_engine.py - agent_utilities/server/concurrency.py doc: FileWatcher — watchdog-triggered graph execution - id: OS-5.1 name: refactoring. This module re-exports it to avoid breaking pillar: OS-5 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/capabilities/stuck_loop.py - agent_utilities/core/config.py - agent_utilities/core/decorators.py - agent_utilities/core/exceptions.py - agent_utilities/decorators.py - agent_utilities/graph/reactive/budget.py - agent_utilities/graph/state.py - agent_utilities/harness/tracing.py - agent_utilities/knowledge_graph/core/owl_bridge.py - agent_utilities/mcp/kg_server.py - agent_utilities/models/knowledge_graph.py - agent_utilities/models/schema_definition.py - agent_utilities/observability/__init__.py - agent_utilities/observability/approval_manager.py - agent_utilities/observability/audit_logger.py - agent_utilities/observability/custom_observability.py - agent_utilities/observability/mermaid.py - agent_utilities/observability/token_tracker.py - agent_utilities/policies/__init__.py - agent_utilities/protocols/a2a_config.py - agent_utilities/security/__init__.py - agent_utilities/security/auth.py - agent_utilities/security/browser_auth.py - agent_utilities/security/cli.py - agent_utilities/security/execution_stability_engine.py - agent_utilities/security/guardrails.py - agent_utilities/security/permissions_kernel.py - agent_utilities/security/secrets_client.py - agent_utilities/security/threat_defense_engine.py - agent_utilities/security/tool_guard.py - agent_utilities/security/xai_auth.py - agent_utilities/server/dependencies.py doc: refactoring. This module re-exports it to avoid breaking - id: OS-5.2 name: MaintenanceCron — scheduled autonomous maintenance pillar: OS-5 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/automation/__init__.py - agent_utilities/automation/file_watcher.py - agent_utilities/automation/maintenance_cron.py - agent_utilities/automation/research_pipeline.py - agent_utilities/core/__init__.py - agent_utilities/core/cognitive_scheduler.py - agent_utilities/core/config.py - agent_utilities/core/registry/package_adapter.py - agent_utilities/core/resource_optimizer.py - agent_utilities/core/scheduler.py - agent_utilities/graph/executor.py - agent_utilities/graph/hierarchical_planner.py - agent_utilities/graph/state.py - agent_utilities/models/knowledge_graph.py - agent_utilities/security/permissions_kernel.py - agent_utilities/security/tool_guard.py - agent_utilities/tools/scheduler_tools.py - agent_utilities/tools/tool_registry.py doc: MaintenanceCron — scheduled autonomous maintenance - id: OS-5.3 name: Reactive Multi-Axis Budget Guardrails pillar: OS-5 status: live code_paths: - agent_utilities/graph/reactive/budget.py doc: Reactive Multi-Axis Budget Guardrails - id: OS-5.4 name: WASM Micro-Agent Sandbox & Runner pillar: OS-5 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/core/__init__.py - agent_utilities/core/wasm_runner.py doc: WASM Micro-Agent Sandbox & Runner - id: OS-5.5 name: Distributed Coordinator with Semantic Sharding pillar: OS-5 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/orchestration/distributed_coordinator.py - agent_utilities/orchestration/recovery_daemon.py - agent_utilities/protocols/a2a.py doc: Distributed Coordinator with Semantic Sharding - id: OS-5.6 name: Deterministic Replay Engine pillar: OS-5 status: live code_paths: - agent_utilities/__init__.py - agent_utilities/observability/replay_engine.py - agent_utilities/security/__init__.py - agent_utilities/security/sandboxed_executor.py doc: Deterministic Replay Engine - id: OS-5.8 name: Epistemic dynamic priority & quota scaling based on KG Centrality pillar: OS-5 status: live code_paths: - agent_utilities/core/cognitive_scheduler.py - agent_utilities/graph/parallel_engine.py - agent_utilities/server/app.py doc: Epistemic dynamic priority & quota scaling based on KG Centrality - id: OS-5.9 name: Gateway Service Dashboard pillar: OS-5 status: live code_paths: - agent_utilities/core/paths.py - agent_utilities/gateway/__init__.py - agent_utilities/gateway/aggregator.py - agent_utilities/gateway/api.py - agent_utilities/gateway/config.py - agent_utilities/gateway/models.py - agent_utilities/gateway/registry.py - agent_utilities/gateway/widgets/base.py - agent_utilities/gateway/ws.py doc: Gateway Service Dashboard - id: OS-5.10 name: Ontological Guardrail Engine pillar: OS-5 status: live code_paths: - agent_utilities/core/sessions.py - agent_utilities/gateway/fleet.py - agent_utilities/gateway/graph_api.py - agent_utilities/security/tool_guard.py doc: Ontological Guardrail Engine - id: OS-5.11 name: establish a run-wide correlation id so every nested pillar: OS-5 status: live code_paths: - agent_utilities/cli/__init__.py - agent_utilities/cli/__main__.py - agent_utilities/core/sessions.py - agent_utilities/gateway/fleet.py - agent_utilities/gateway/fleet_events.py - agent_utilities/observability/correlation.py - agent_utilities/observability/langfuse_exporter.py - agent_utilities/orchestration/agent_dispatch_worker.py - agent_utilities/orchestration/agent_runner.py - agent_utilities/orchestration/engine.py - agent_utilities/security/run_token.py doc: establish a run-wide correlation id so every nested - id: OS-5.12 name: Empty → derive from effect pillar: OS-5 status: live code_paths: - agent_utilities/knowledge_graph/actions/executor.py - agent_utilities/knowledge_graph/actions/models.py - agent_utilities/observability/escalation_matrix.py doc: Empty → derive from effect - id: OS-5.13 name: resolve the active channel once; pillar: OS-5 status: live code_paths: - agent_utilities/core/release_channel.py - agent_utilities/graph/routing/enrichers/capability_designation.py doc: resolve the active channel once; - id: OS-5.14 name: an ACL-check exception denies, and nodes without an ACL are pillar: OS-5 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/core/sessions.py - agent_utilities/gateway/graph_api.py - agent_utilities/knowledge_graph/core/graph_compute.py - agent_utilities/knowledge_graph/distillation/skill_graph_distiller.py - agent_utilities/knowledge_graph/facade.py - agent_utilities/knowledge_graph/ingest_worker.py - agent_utilities/knowledge_graph/ontology/permissioning.py - agent_utilities/mcp/kg_server.py - agent_utilities/observability/correlation.py - agent_utilities/orchestration/agent_dispatch_worker.py - agent_utilities/security/brain_context.py - agent_utilities/security/request_identity.py doc: an ACL-check exception denies, and nodes without an ACL are - id: OS-5.15 name: Fleet Event Ingress pillar: OS-5 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/gateway/fleet.py - agent_utilities/gateway/fleet_events.py - agent_utilities/knowledge_graph/adaptation/failure_analyzer.py - agent_utilities/knowledge_graph/adaptation/fleet_event_triage.py - agent_utilities/knowledge_graph/core/engine_tasks.py doc: Fleet Event Ingress - id: OS-5.16 name: Unified durable-state externalization — one STATE_DB_URI flag selects a shared Postgres state store over the per-host SQLite default pillar: OS-5 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/core/leadership.py - agent_utilities/core/sessions.py - agent_utilities/core/state_store.py - agent_utilities/gateway/fleet.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/queue_backend.py - agent_utilities/mcp/tools/state_tools.py - agent_utilities/orchestration/agent_dispatch.py - agent_utilities/orchestration/agent_dispatch_worker.py - agent_utilities/orchestration/durable_execution.py doc: Unified durable-state externalization — one STATE_DB_URI flag selects a shared Postgres state store over the per-host SQLite default - id: OS-5.17 name: Cross-host daemon leadership election via Postgres advisory locks so singleton background ticks run on exactly one host fleet-wide pillar: OS-5 status: live code_paths: - agent_utilities/core/leadership.py - agent_utilities/knowledge_graph/core/engine_tasks.py doc: Cross-host daemon leadership election via Postgres advisory locks so singleton background ticks run on exactly one host fleet-wide - id: OS-5.18 name: Fleet supervisory plane at scale — SQL aggregation, paginated and filtered session queries, and desired-state pause and kill reconciliation across hosts pillar: OS-5 status: live code_paths: - agent_utilities/gateway/fleet.py doc: Fleet supervisory plane at scale — SQL aggregation, paginated and filtered session queries, and desired-state pause and kill reconciliation across hosts - id: OS-5.23 name: No-op without the metrics extra pillar: OS-5 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/gateway/api.py - agent_utilities/gateway/config.py - agent_utilities/gateway/graph_api.py - agent_utilities/gateway/rate_limit.py - agent_utilities/knowledge_graph/core/engine_breaker.py - agent_utilities/knowledge_graph/core/graph_compute.py - agent_utilities/mcp/middlewares.py - agent_utilities/mcp/server_factory.py - agent_utilities/observability/gateway_metrics.py - agent_utilities/orchestration/scaling_signals.py - agent_utilities/security/request_identity.py - agent_utilities/server/__init__.py doc: No-op without the metrics extra - id: OS-5.24 name: Operational ActionPolicy decision point with per-action autonomy tiers, durable rate limits, maintenance windows, and blast-radius caps consulted before any autonomous mutating fleet action pillar: OS-5 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/gateway/fleet.py - agent_utilities/knowledge_graph/adaptation/remediation_playbooks.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/research/auto_merge.py - agent_utilities/knowledge_graph/research/change_publisher.py - agent_utilities/orchestration/action_policy.py - agent_utilities/orchestration/fleet_actuation.py - agent_utilities/orchestration/fleet_autoscaler.py - agent_utilities/orchestration/fleet_reconciler.py - agent_utilities/runtime/policy.py doc: Operational ActionPolicy decision point with per-action autonomy tiers, durable rate limits, maintenance windows, and blast-radius caps consulted before any autonomous mutating fleet action - id: OS-5.25 name: Desired-state fleet reconciler pillar: OS-5 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/gateway/fleet.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/research/change_publisher.py - agent_utilities/orchestration/action_policy.py - agent_utilities/orchestration/fleet_actuation.py - agent_utilities/orchestration/fleet_observation.py - agent_utilities/orchestration/fleet_reconciler.py doc: Desired-state fleet reconciler - id: OS-5.26 name: Remediation playbooks with stepwise verification and policy-gated actuation plugged into the fleet-event triage seam, recording every step on the originating FleetEvent node pillar: OS-5 status: live code_paths: - agent_utilities/knowledge_graph/adaptation/remediation_playbooks.py - agent_utilities/knowledge_graph/core/engine_tasks.py doc: Remediation playbooks with stepwise verification and policy-gated actuation plugged into the fleet-event triage seam, recording every step on the originating FleetEvent node - id: OS-5.27 name: Health-gated deploy and rollback watch where every deploy or restart the autonomy plane triggers is followed by a durable health watch whose failure invokes a policy-gated rollback and escalation pillar: OS-5 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/orchestration/deploy_watch.py doc: Health-gated deploy and rollback watch where every deploy or restart the autonomy plane triggers is followed by a durable health watch whose failure invokes a policy-gated rollback and escalation - id: OS-5.28 name: Shard topology visibility with per-shard reachability status surfaces and per-endpoint engine gauges and counters pillar: OS-5 status: live code_paths: - agent_utilities/gateway/api.py - agent_utilities/knowledge_graph/core/engine_breaker.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/knowledge_graph/core/shard_topology.py - agent_utilities/mcp/kg_server.py doc: Shard topology visibility with per-shard reachability status surfaces and per-endpoint engine gauges and counters - id: OS-5.29 name: Reactive replica autoscaling pillar: OS-5 status: live code_paths: - agent_utilities/core/config.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/orchestration/action_policy.py - agent_utilities/orchestration/fleet_autoscaler.py - agent_utilities/orchestration/fleet_reconciler.py - agent_utilities/orchestration/scaling_signals.py doc: Reactive replica autoscaling - id: OS-5.30 name: Declarative skill / skill-workflow scheduler pillar: OS-5 status: live code_paths: - agent_utilities/core/skill_scheduler.py - agent_utilities/knowledge_graph/core/engine_tasks.py - agent_utilities/server/routers/commands.py doc: Declarative skill / skill-workflow scheduler - id: OS-5.31 name: persist this graph run as a runtime usage row so pillar: OS-5 status: live code_paths: - agent_utilities/observability/gateway_metrics.py - agent_utilities/orchestration/engine.py - agent_utilities/usage/recorder.py doc: persist this graph run as a runtime usage row so - id: OS-5.32 name: so jwt-protected children are reachable, then pillar: OS-5 status: live code_paths: - agent_utilities/knowledge_graph/core/engine_mcp_discovery.py - agent_utilities/mcp/child_resilience.py - agent_utilities/mcp/client_credentials.py - agent_utilities/mcp/multiplexer.py doc: so jwt-protected children are reachable, then - id: OS-5.33 name: Bridge the developer-workspace mutating-action gate to the fleet pillar: OS-5 status: live code_paths: - agent_utilities/orchestration/action_policy.py - agent_utilities/runtime/docker_workspace.py - agent_utilities/runtime/local_workspace.py - agent_utilities/runtime/policy.py - agent_utilities/runtime/workspace.py - agent_utilities/server/app.py - agent_utilities/server/routers/runtime.py doc: Bridge the developer-workspace mutating-action gate to the fleet - id: OS-5.34 name: the KG-provenance panel data for a run pillar: OS-5 status: live code_paths: - agent_utilities/server/routers/runtime.py doc: the KG-provenance panel data for a run - id: OS-5.35 name: cost-and-throughput aware autoscaling that caps scale-up to a configured hourly budget and surfaces a throughput-per-dollar efficiency metric instead of reacting to a single load signal alone pillar: OS-5 status: live code_paths: - agent_utilities/orchestration/cost_governor.py - agent_utilities/orchestration/fleet_autoscaler.py doc: cost-and-throughput aware autoscaling that caps scale-up to a configured hourly budget and surfaces a throughput-per-dollar efficiency metric instead of reacting to a single load signal alone - id: OS-5.36 name: a search-distillation harvester that converts the reasoning router pillar: OS-5 status: live code_paths: - agent_utilities/harness/search_distillation.py - agent_utilities/knowledge_graph/core/reasoner.py doc: a search-distillation harvester that converts the reasoning router - id: OS-5.37 name: a value-aware test-time-compute governor that decides whether spending more reasoning compute is worth it by stopping once a result satisfices or the marginal expected quality gain per extra attempt falls below a threshold pillar: OS-5 status: live code_paths: - agent_utilities/harness/compute_governor.py - agent_utilities/knowledge_graph/core/reasoner.py doc: a value-aware test-time-compute governor that decides whether spending more reasoning compute is worth it by stopping once a result satisfices or the marginal expected quality gain per extra attempt falls below a threshold - id: OS-5.38 name: Universal Outbound CredentialProvider pillar: OS-5 status: live code_paths: - agent_utilities/security/__init__.py - agent_utilities/security/credential_provider.py - agent_utilities/security/source_credentials.py doc: Universal Outbound CredentialProvider - id: OS-5.39 name: Typed Source-Credential Registry pillar: OS-5 status: live code_paths: - agent_utilities/security/__init__.py - agent_utilities/security/source_credentials.py doc: Typed Source-Credential Registry - id: OS-5.40 name: Governance-derived Claude Code permission-fence generator pillar: OS-5 status: live code_paths: - agent_utilities/claude_harness/__init__.py - agent_utilities/claude_harness/claude_fence.py - agent_utilities/cli/__init__.py - agent_utilities/deployment/cli.py - agent_utilities/mcp/tools/analysis_tools.py doc: Governance-derived Claude Code permission-fence generator - id: OS-5.41 name: Dynamic two-layer fail-closed PreToolUse ActionPolicy permission gate pillar: OS-5 status: live code_paths: - agent_utilities/claude_harness/__init__.py - agent_utilities/claude_harness/pretooluse_gate.py - agent_utilities/cli/__init__.py - agent_utilities/ecosystem/hook_installer.py - agent_utilities/orchestration/action_policy.py doc: Dynamic two-layer fail-closed PreToolUse ActionPolicy permission gate - id: OS-5.42 name: Multi-session concept-ID allocation & coordination protocol pillar: OS-5 status: live code_paths: - agent_utilities/cli/__init__.py - agent_utilities/governance/concept_allocator.py - agent_utilities/mcp/tools/ontology_tools.py doc: Multi-session concept-ID allocation & coordination protocol - id: SAFE-1.0 name: Tool-Agnostic File Safety Hooks pillar: SAFE-1 status: live code_paths: - agent_utilities/file_safety.py doc: Tool-Agnostic File Safety Hooks - id: SAFE-1.1 name: non-saturating superhuman progress tracking via relative scorers and a saturation detector that keep producing signal past the human or known-answer ceiling so a genuine capability jump is distinguishable from metric saturation pillar: SAFE-1 status: live code_paths: - agent_utilities/harness/frontier_scorers.py - agent_utilities/harness/reliability_scorers.py doc: non-saturating superhuman progress tracking via relative scorers and a saturation detector that keep producing signal past the human or known-answer ceiling so a genuine capability jump is distinguishable from metric saturation - id: SAFE-1.2 name: a multi-agent scaling-law harness that sweeps collective size over a fixed task and fits capability ~ N^alpha so the platform can measure whether adding agents helps super- or sub-linearly instead of assuming it does pillar: SAFE-1 status: live code_paths: - agent_utilities/harness/scaling_laws.py doc: a multi-agent scaling-law harness that sweeps collective size over a fixed task and fits capability ~ N^alpha so the platform can measure whether adding agents helps super- or sub-linearly instead of assuming it does - id: SAFE-1.3 name: recursive-improvement velocity tracker that surfaces whether the loop is still improving and flags a non-positive derivative as a research-gets-harder signal pillar: SAFE-1 status: live code_paths: - agent_utilities/knowledge_graph/research/improvement_ledger.py doc: recursive-improvement velocity tracker that surfaces whether the loop is still improving and flags a non-positive derivative as a research-gets-harder signal - id: SAFE-1.4 name: a model-collapse guard for the self-generated training corpus that rejects near-duplicate or distributionally-narrowing rows and caps the synthetic-to-human fraction so recursive distillation cannot quietly degenerate pillar: SAFE-1 status: live code_paths: - agent_utilities/harness/corpus_collapse_guard.py doc: a model-collapse guard for the self-generated training corpus that rejects near-duplicate or distributionally-narrowing rows and caps the synthetic-to-human fraction so recursive distillation cannot quietly degenerate - id: SAFE-1.5 name: objective-level safety primitives for rising autonomy pillar: SAFE-1 status: live code_paths: - agent_utilities/core/corrigibility.py - agent_utilities/orchestration/action_policy.py doc: objective-level safety primitives for rising autonomy - id: SAFE-1.6 name: SAFE-1.6 pillar: SAFE-1 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/superhuman_gate.py doc: SAFE-1.6 - id: SAFE-1.7 name: SAFE-1.7 pillar: SAFE-1 status: live code_paths: - agent_utilities/harness/__init__.py - agent_utilities/harness/adaptation_benchmark.py doc: SAFE-1.7 - id: SAFE-1.8 name: Unattended-session stop-on-ask containment pillar: SAFE-1 status: live code_paths: - agent_utilities/claude_harness/overnight_runner.py doc: Unattended-session stop-on-ask containment - id: UTIL-1.0 name: Data Type Conversion pillar: UTIL-1 status: live code_paths: - agent_utilities/base_utilities.py doc: Data Type Conversion