Orchestrata Ultimate / RTM / 107-context-injection-improvements
RTM: 107-context-injection-improvements
Requirement to symbol to test rows for 107-context-injection-improvements: Context-Injection Improvements.
At a glance
| Item | Value |
|---|---|
| Requirement IDs | 73 |
Traced from src/ |
13 (17.8%) |
| Untraced | 60 |
Cited from tests/ |
1 (1.4%) |
Traced in src/ but not cited from tests/ |
12 |
Reading this matrix
This matrix lists each of the 73 requirement IDs of spec 107-context-injection-improvements with the src/ symbols that cite it and the tests that cite it. 13 of them (17.8%) have at least one citing symbol and 60 have none.
1 requirement is also cited from tests/, and 12 are traced in code but never named by a test. A test citation is a @traces on a test function; it shows intent to cover, not that the test runs or reaches the symbol.
Rows showing none in the symbols column are the gaps to close first. The densest untraced groups are REQ-107-* (60).
Requirement to symbols to tests
| Requirement | Symbols in src/ |
Tests citing it |
|---|---|---|
REQ-107-001 Production PipelineRunner and gateway paths load ContextData from the active... |
none | none |
REQ-107-002 mission_context.render_mission_context() is rendered into the mission layer when... |
none | none |
REQ-107-003 A plan_context provider renders recent pipeline outcomes / next planned stages into... |
orchestrata_ultimate/context/providers/plan_context render_plan_context L20orchestrata_ultimate/planning/plan_parser parse_plan_file L228orchestrata_ultimate/planning/plan_parser parse_plan_text L194 |
tests/knowledge_graph/test_graph_context_validator.py::test_extracts_requirement_ids |
REQ-107-004 An observed test proves the mission layer is non-empty when a session has active goals |
none | none |
REQ-107-060 PlanParser parses plans/execution-plan.md into structured `(step_num, title, goal,... |
orchestrata_ultimate/planning/plan_parser _extract_step_body_fields L94orchestrata_ultimate/planning/plan_parser _parse_actions L132orchestrata_ultimate/planning/plan_parser _parse_phase_steps L154+5 more |
none |
REQ-107-061 Goal persists a current_step index; the session controller provides... |
none | none |
REQ-107-062 plan_context provider uses current_step to render the current step in full plus... |
none | none |
REQ-107-005 MemoryEngine.search accepts a max_entries pre-filter before similarity scoring |
none | none |
REQ-107-006 MemoryEngine.search accepts a max_age_seconds pre-filter before similarity scoring |
none | none |
REQ-107-007 Default filters (max_entries=50, max_age_seconds=3600) keep memory recall bounded... |
none | none |
REQ-107-008 An observed test proves recall returns at most max_entries results |
none | none |
REQ-107-009 SemanticGraphContextProvider queries Memgraph for chunks relevant to a query string |
none | none |
REQ-107-010 The provider contributes a semantic part to ContextAssembly for subjects... |
none | none |
REQ-107-011 When Memgraph is unavailable the provider returns an empty string without raising |
none | none |
REQ-107-012 An observed test proves semantic context is injected for implement subject |
none | none |
REQ-107-013 AST complexity, branch, and call-graph signals are queryable from Neo4j or the AST... |
none | none |
REQ-107-014 An ast_context provider returns a budgeted summary of the most relevant... |
orchestrata_ultimate/context/providers/ast_context render_ast_context L106 |
none |
REQ-107-015 When the AST/KG source is unavailable the provider returns an empty string without raising |
orchestrata_ultimate/context/providers/semantic_context render_semantic_context L50orchestrata_ultimate/knowledge_graph/context_injector _query_semantic_bridge L1211orchestrata_ultimate/knowledge_graph/semantic_bridge render_semantic_roots L238+1 more |
none |
REQ-107-016 An observed test proves ast_context is non-empty for a task targeting a known module |
none | none |
REQ-107-020 Every UnifiedContextAssembler.assemble() call logs ContextUsage.to_dict() at INFO |
orchestrata_ultimate/context/usage_log write_usage_record L37 |
none |
REQ-107-021 scripts/context_usage_report.py exists and can aggregate per-layer token spend from... |
orchestrata_ultimate/context/usage_log read_usage_records L74 |
none |
REQ-107-022 The report script can aggregate per-volk and per-subject token spend |
none | none |
REQ-107-030 ArtifactStore accepts an optional output_dir and persists each artifact to a JSONL... |
none | none |
REQ-107-031 ArtifactStore replays the JSONL file into the in-memory index on construction |
none | none |
REQ-107-032 render_artifact_digest ranks artifacts by recency/relevance and renders only top-N... |
orchestrata_ultimate/core/artifact_store render_artifact_digest L392 |
none |
REQ-107-033 PipelineArtifact carries optional spec and reqs fields |
none | none |
REQ-107-034 ArtifactStore.by_requirement(req_id) returns artifacts whose reqs include that id |
none | none |
REQ-107-035 An observed test proves artifacts survive an ArtifactStore reconstruct cycle |
none | none |
REQ-107-050 get_code tool returns content or signature for a canonical code entity id |
orchestrata_ultimate/tools/context_tools GetCodeTool L204orchestrata_ultimate/tools/context_tools _resolve_by_path_and_name L292orchestrata_ultimate/tools/context_tools _split_entity_id L55+1 more |
none |
REQ-107-051 get_tests tool returns covering tests for a canonical code entity |
orchestrata_ultimate/tools/context_tools GetTestsTool L348orchestrata_ultimate/tools/context_tools _resolve_tests_fallback L425orchestrata_ultimate/tools/context_tools _split_entity_id L55+1 more |
none |
REQ-107-052 get_requirements tool returns traced requirements for a canonical code entity |
orchestrata_ultimate/tools/context_tools GetRequirementsTool L461orchestrata_ultimate/tools/context_tools _collect_fallback_item_ids L569orchestrata_ultimate/tools/context_tools _collect_traced_item_ids L543+2 more |
none |
REQ-107-053 search_code tool returns semantic-search context for a natural-language query |
orchestrata_ultimate/tools/context_tools SearchCodeTool L616orchestrata_ultimate/tools/context_tools execute L644 |
none |
REQ-107-054 Tools degrade gracefully when Neo4j or Memgraph is unavailable |
orchestrata_ultimate/tools/context_tools GetCodeTool L204orchestrata_ultimate/tools/context_tools GetRequirementsTool L461orchestrata_ultimate/tools/context_tools GetTestsTool L348+8 more |
none |
REQ-107-055 Tools are registered as SPEC-062 native tools and exposed to Architect, CodeDirector,... |
orchestrata_ultimate/tools/context_tools register_context_tools L718 |
none |
REQ-107-056 Observed tests prove each tool against real Neo4j/Memgraph |
none | none |
REQ-107-070 UnifiedContextAssembler caches rendered ContextPack output by `(snapshot_id,... |
none | none |
REQ-107-071 An observed test proves the cache returns identical context parts for repeated calls... |
none | none |
REQ-107-080 GraphContext populates a recent_changes section from explicit Change nodes,... |
none | none |
REQ-107-081 GraphContext populates a similar_patterns section with files sharing base classes... |
none | none |
REQ-107-082 GraphContext populates a semantic_files section via the Memgraph semantic-search... |
none | none |
REQ-107-083 GraphContext populates requirements and decisions sections by traversing... |
none | none |
REQ-107-084 Requirement closure walks REQUIRES edges up to a bounded depth and returns... |
none | none |
REQ-107-090 Runtime signals (Message, ToolCall, Decision, GatewayEvent) are persisted to... |
none | none |
REQ-107-091 Runtime signals link to mentioned File/Class/Function nodes via MENTIONS edges |
none | none |
REQ-107-092 Kernel pipeline dispatch, orchestrator runner, and BaseAgent wire the... |
none | none |
REQ-107-100 GraphContextValidator validates LLM responses against core entities cited in... |
none | none |
REQ-107-101 ValidatorLoop retries LLM calls with expanded context when validation fails |
none | none |
REQ-107-102 ContextExpansion controls extra_hops, include_tested_by,... |
none | none |
REQ-107-110 ast-sync passes --incremental, warm-starting from the snapshot on disk and... |
none | none |
REQ-107-111 ast-sync --snapshot-only runs steps 1/1b and stops, skipping the frozen-monorepo... |
none | none |
REQ-107-112 main() parses the argv it is given; argv=None inheriting sys.argv aborts an... |
none | none |
REQ-107-113 Freshness is evaluated at the point of USE and scoped to the files the task touches,... |
none | none |
REQ-107-114 A consumer that cannot reach the KB distinguishes "queried, answer is nothing" from... |
none | none |
REQ-107-115 Blast radius resolves impact targets from the module's parsed AST, not module-api... |
none | none |
REQ-107-116 Every bound on an answer (fan-out, parse caps, project scoping) is stated in the output |
none | none |
REQ-107-117 code_intel.py impact accepts many qnames in one invocation, and the six-symbol... |
none | none |
REQ-107-120 claude/ in this repo is the source of truth; ~/.claude is a deployment of it and... |
none | none |
REQ-107-121 The deployment copies files — no symlink from ~/.claude into a checkout |
none | none |
REQ-107-122 The deployment owns a uv venv on python3.14t, and ~/.claude/.envrc... |
none | none |
REQ-107-123 The installer rewrites settings.json hook commands to the venv's absolute... |
none | none |
REQ-107-124 The installer records the source commit and refuses success unless every deployed file... |
none | none |
REQ-107-125 The drift comparison survives SIGPIPE and trailing-newline stripping — both previously... |
none | none |
REQ-107-126 Deployed files are enumerated, never selected by filename heuristic... |
none | none |
REQ-107-127 A verification that matched zero files fails (CHECKED -gt 0 asserted explicitly) |
none | none |
REQ-107-128 The deployed venv is verified free-threaded, and no deployed file references the... |
none | none |
REQ-107-130 ship runs verify → push → deploy → verify-the-deployment in one sequence |
none | none |
REQ-107-131 ship is reachable by name on PATH, and the installer verifies that by name rather... |
none | none |
REQ-107-132 Each project states its own gate in ship; an unknown project fails rather than... |
none | none |
REQ-107-133 The dirty-tree check excludes tooling-rewritten paths, narrowly — sibling hand-written... |
none | none |
REQ-107-134 A Stop hook refuses to end a turn while a watched repo holds local-only commits |
none | none |
REQ-107-135 That hook stays quiet while a ship for the repo is already running, detecting it by... |
none | none |
REQ-107-136 In-flight detection is by process, not lockfile — a lockfile outliving a SIGKILL would... |
none | none |
REQ-107-137 --no-verify states that it skips only this script's gate, and warns that an... |
none | none |