Summary

• switch the Koru semantic judge from stale DeepSeek to benchmark-qualified google/gemini-3.1-pro-preview
• keep Vallm semantic/security review fail-closed while leaving TypeScript regression to the required verify job
• record commit-bound no-source and live two-file review evidence

Local validation

• npm run verify:workflows — PASS
• npm run verify — PASS (335 tests, 334 passed, 1 JDK-dependent skip)
• focused model tests — PASS 2/2
• Vallm compatibility probe — deterministic/security PASS 2/2 after lowercase language normalization
• make governance — expected inherited ticket-019 findings only; no ticket-018 scope/secret finding

Timeout and failure policy

• complete GitHub job: 10 minutes
• Vallm/LiteLLM review: 420 seconds; response cap: 8192 tokens
• provider retries forced to zero, so HTTP errors including 404 fail immediately
• Vallm 0.1.94 TypeScript language ID normalized before tree-sitter parsing
• semantic info/warnings are advisory only with a passing file verdict; semantic errors and all deterministic, provider, malformed/missing-result and timeout findings remain blocking and visible in JSON

Live evidence

Repository-scoped OPENROUTER_API_KEY was rotated through gh stdin after explicit user authorization. Dispatch 30714664770 reached openrouter/google/gemini-3.1-pro-preview for both PR #3 TypeScript files without authentication or provider-limit errors. Both file-level verdicts were pass; that run exposed the parser and aggregate-policy defects repaired by this PR.

Repaired live evidence

Dispatch 30746421293 reviewed PR #3 range 2e87205..6b79527 in 1m24s. The attested report records Gemini, both expected TypeScript files, verdict pass, exit 0, 2/2 passed, zero blocking/parser/provider findings, and five retained advisory semantic observations. Vallm original exit 2 remains recorded before policy normalization.

Remaining blockers

• inherited ticket-019 governance conflicts
• independent protected review/merge evidence

Follow-up failure-path validation

• actual LiteLLM HTTP stack + local 404 endpoint: in ~705 ms, exactly one request,
• slow local endpoint + 0.5 s request limit: in ~799 ms, exactly one request
• fresh

todo2code@0.5.1 verify
npm run check && npm run verify:no-llm && npm run verify:modules && npm run verify:env && npm run verify:workflows && npm run verify:generated-analysis && npm run verify:structured-responses && npm run build && npm run verify:schemas && npm test

todo2code@0.5.1 check
tsc -p tsconfig.json --noEmit

todo2code@0.5.1 verify:no-llm
node scripts/verify-no-llm-imports.mjs

LLM boundary verified transitively from 9 deterministic entrypoints across 38 modules.

todo2code@0.5.1 verify:modules
node scripts/verify-module-boundaries.mjs

Module boundaries verified: 114 modules, 521 internal imports, no cycles, core is independent.

todo2code@0.5.1 verify:env
node scripts/verify-env-contract.mjs

Environment contract verified: 75 code/Docker variables, 75 documented keys, no duplicates.

todo2code@0.5.1 verify:workflows
node scripts/verify-workflow-yaml.mjs

Workflow YAML verified: 2 file(s), no duplicate top-level keys.

todo2code@0.5.1 verify:generated-analysis
node scripts/verify-generated-analysis.mjs

{"filesChecked":19,"untrackedInputsChecked":0,"status":"ok"}

todo2code@0.5.1 verify:structured-responses
node scripts/verify-structured-responses.mjs

{"structuredCalls":7,"rawCalls":0,"status":"ok"}

todo2code@0.5.1 build
tsc -p tsconfig.json

todo2code@0.5.1 verify:schemas
node scripts/generate-response-schemas.mjs --check

{"schema":"schemas/document-extraction-response.schema.json","status":"ok"}

todo2code@0.5.1 test
node --test --test-concurrency=4 dist/test/*.test.js

TAP version 13

[t2c:a2a] listening on 127.0.0.1:38195

[t2c:a2a] listening on 127.0.0.1:37603

Subtest: A2A governed-intake skill completes accepted commands and rejects domain failures

ok 1 - A2A governed-intake skill completes accepted commands and rejects domain failures

duration_ms: 133.049905
...

[t2c:a2a] listening on 127.0.0.1:35497

Subtest: A2A v1.0 card, versioning, task methods and cursor pagination are coherent

ok 2 - A2A v1.0 card, versioning, task methods and cursor pagination are coherent

duration_ms: 140.624917
...

[t2c:a2a] listening on 127.0.0.1:41177

Subtest: A2A bearer authentication is declared with v1 security objects and enforced

ok 3 - A2A bearer authentication is declared with v1 security objects and enforced

duration_ms: 32.579244
...

[t2c:a2a] listening on 127.0.0.1:41259

[t2c:a2a] listening on 127.0.0.1:32825

[t2c:a2a] listening on 127.0.0.1:43233

Subtest: A2A file task store survives restart and preserves idempotency across replicas

ok 4 - A2A file task store survives restart and preserves idempotency across replicas

duration_ms: 110.351699
...

Subtest: the A2A bind default is loopback and a wildcard bind is not mistaken for it

ok 5 - the A2A bind default is loopback and a wildcard bind is not mistaken for it

duration_ms: 1.834447
...

Subtest: Go adapter records package, imports, types, functions and methods

ok 6 - Go adapter records package, imports, types, functions and methods

duration_ms: 191.156694
...

Subtest: Go facts are deterministic observations, not inferences

ok 7 - Go facts are deterministic observations, not inferences

duration_ms: 144.172903
...

Subtest: Go adapter marks exported symbols and reports calls in scope

ok 8 - Go adapter marks exported symbols and reports calls in scope

duration_ms: 113.207635
...

Subtest: Go extraction is skipped without cost when a tree holds no Go sources

ok 9 - Go extraction is skipped without cost when a tree holds no Go sources

duration_ms: 14.56097
...

Subtest: A missing Go toolchain degrades to a warning instead of failing the run

ok 10 - A missing Go toolchain degrades to a warning instead of failing the run

duration_ms: 11.266279
...

Subtest: Rust adapter records uses, types, functions, methods, values and calls

ok 11 - Rust adapter records uses, types, functions, methods, values and calls

duration_ms: 104.750893
...

Subtest: Java adapter records packages, imports, types, fields, methods and calls

ok 12 - Java adapter records packages, imports, types, fields, methods and calls # SKIP JDK not installed

duration_ms: 8.966518
...

Subtest: Java and Rust adapters skip toolchain startup when no matching sources exist

ok 13 - Java and Rust adapters skip toolchain startup when no matching sources exist

duration_ms: 21.042575
...

Subtest: Missing Java and Rust toolchains degrade to explicit warnings

ok 14 - Missing Java and Rust toolchains degrade to explicit warnings

duration_ms: 51.660978
...

Subtest: PHP syntax adapter records namespaces, imports, types, functions, methods and calls

ok 15 - PHP syntax adapter records namespaces, imports, types, functions, methods and calls

duration_ms: 101.635199
...

Subtest: PHP adapter skips runtime startup when no PHP source exists

ok 16 - PHP adapter skips runtime startup when no PHP source exists

duration_ms: 16.947268
...

Subtest: Missing PHP runtime degrades to an explicit warning

ok 17 - Missing PHP runtime degrades to an explicit warning

duration_ms: 9.744386
...

Subtest: Invalid PHP syntax is reported without aborting extraction

ok 18 - Invalid PHP syntax is reported without aborting extraction

duration_ms: 75.710835
...

Subtest: AST extractor reads TypeScript and Python facts

ok 19 - AST extractor reads TypeScript and Python facts

duration_ms: 110.625447
...

Subtest: CLI command help is successful and non-mutating

ok 20 - CLI command help is successful and non-mutating

duration_ms: 2375.182123
...

Subtest: TypeScript and Python CLIs execute the same intake command/query handler

ok 21 - TypeScript and Python CLIs execute the same intake command/query handler

duration_ms: 1629.798729
...

Subtest: CLI summarize exposes deterministic, prefer-llm and require-llm modes

ok 22 - CLI summarize exposes deterministic, prefer-llm and require-llm modes

duration_ms: 4100.87357
...

Subtest: CLI propose-todo, render-todo and apply-todo return JSON and preserve a no-op TODO

ok 23 - CLI propose-todo, render-todo and apply-todo return JSON and preserve a no-op TODO

duration_ms: 3147.502079
...

Subtest: CLI watch reads TASK.md by default, disables summary LLM and reacts to a live file change

ok 24 - CLI watch reads TASK.md by default, disables summary LLM and reacts to a live file change

duration_ms: 2086.17955
...

Subtest: proposeCodeChangePlans materialises grounded plans from PLANNED_NOT_IMPLEMENTED

ok 25 - proposeCodeChangePlans materialises grounded plans from PLANNED_NOT_IMPLEMENTED

duration_ms: 31.288063
...

Subtest: code-change title preserves the leading action of a compound intent

ok 26 - code-change title preserves the leading action of a compound intent

duration_ms: 2.31415
...

Subtest: proposeCodeChangePlans is deterministic for the same evidence

ok 27 - proposeCodeChangePlans is deterministic for the same evidence

duration_ms: 3.980339
...

Subtest: a plan creates a missing file and modifies an existing one

ok 28 - a plan creates a missing file and modifies an existing one

duration_ms: 11.564583
...

Subtest: a plan never invents a repository-root file from a bare filename

ok 29 - a plan never invents a repository-root file from a bare filename

duration_ms: 3.953313
...

Subtest: a plan whose every path is bare and missing is withheld entirely

ok 30 - a plan whose every path is bare and missing is withheld entirely

duration_ms: 1.724869
...

Subtest: a plan never targets a location outside the repository tree

ok 31 - a plan never targets a location outside the repository tree

duration_ms: 0.312334
...

Subtest: the repository probe never proposes creating a file outside the root

ok 32 - the repository probe never proposes creating a file outside the root

duration_ms: 1.263122
...

Subtest: bounded plan sets prefer explicit TODO work over historical changelog audit

ok 33 - bounded plan sets prefer explicit TODO work over historical changelog audit

duration_ms: 3.550929
...

Subtest: evaluateCodeChangeAcceptance passes when targeted diagnostics clear

ok 34 - evaluateCodeChangeAcceptance passes when targeted diagnostics clear

duration_ms: 5.921532
...

Subtest: evaluateCodeChangeAcceptance fails while the plan is still open

ok 35 - evaluateCodeChangeAcceptance fails while the plan is still open

duration_ms: 1.819317
...

Subtest: Plans without repository paths are not invented

ok 36 - Plans without repository paths are not invented

duration_ms: 0.986083
...

Subtest: Non-repository paths are ignored instead of aborting code-change planning

ok 37 - Non-repository paths are ignored instead of aborting code-change planning

duration_ms: 1.280131
...

Subtest: Acceptance rejects ungrounded paths, missing provenance and inconsistent verdicts

ok 38 - Acceptance rejects ungrounded paths, missing provenance and inconsistent verdicts

duration_ms: 4.742995
...

Subtest: createCodeChangeSourcePatch is deterministic and path-bound

ok 39 - createCodeChangeSourcePatch is deterministic and path-bound

duration_ms: 4.745655
...

Subtest: applyUnifiedDiffToText creates and modifies files from hunks

ok 40 - applyUnifiedDiffToText creates and modifies files from hunks

duration_ms: 0.728403
...

Subtest: applyCodeChangeSourcePatch requires approval and is idempotent

ok 41 - applyCodeChangeSourcePatch requires approval and is idempotent

duration_ms: 14.276191
...

Subtest: applyCodeChangeSourcePatch preflights diffs and refuses symlink escapes

ok 42 - applyCodeChangeSourcePatch preflights diffs and refuses symlink escapes

duration_ms: 5.517787
...

Subtest: createCodeChangeSourcePatchSet covers every plan

ok 43 - createCodeChangeSourcePatchSet covers every plan

duration_ms: 2.901765
...

Subtest: createCodeChangeReviewPatch is hash-stable and lists grounded paths

ok 44 - createCodeChangeReviewPatch is hash-stable and lists grounded paths

duration_ms: 2.789541
...

Subtest: CLI proposes and evaluates a grounded code-change plan through persisted JSON

ok 45 - CLI proposes and evaluates a grounded code-change plan through persisted JSON

duration_ms: 3226.923974
...

Subtest: isUsefulCodeChangePath rejects vendored, binary and analysis dump paths

ok 46 - isUsefulCodeChangePath rejects vendored, binary and analysis dump paths

duration_ms: 0.475118
...

Subtest: proposeCodeChangePlans skips diagnostics that only name junk paths

ok 47 - proposeCodeChangePlans skips diagnostics that only name junk paths

duration_ms: 0.925965
...

Subtest: Published code-change JSON schemas require provenance, risk and rollback

ok 48 - Published code-change JSON schemas require provenance, risk and rollback

duration_ms: 1.832967
...

Subtest: participant registry maps stable IDs to Git/A2A identifiers without display-name guessing

ok 49 - participant registry maps stable IDs to Git/A2A identifiers without display-name guessing

duration_ms: 52.365262
...

Subtest: participant registry rejects ambiguous external identifiers

ok 50 - participant registry rejects ambiguous external identifiers

duration_ms: 0.672329
...

Subtest: trusted intake persists roles across tickets and fails closed without rejected writes

ok 51 - trusted intake persists roles across tickets and fails closed without rejected writes

duration_ms: 117.61495
...

Subtest: agents cannot create human projections and a damaged event hash fails verification

ok 52 - agents cannot create human projections and a damaged event hash fails verification

duration_ms: 25.243968
...

Subtest: TypeScript and dependency-free Python codecs share golden bytes and preserve unknown fields

ok 53 - TypeScript and dependency-free Python codecs share golden bytes and preserve unknown fields

duration_ms: 80.695325
...

Subtest: communication enrichment preserves runtime identity, source, ticket and epistemic class

ok 54 - communication enrichment preserves runtime identity, source, ticket and epistemic class

duration_ms: 66.589084
...

Subtest: communication enrichment corrects one rejected structured response without weakening validation

ok 55 - communication enrichment corrects one rejected structured response without weakening validation

duration_ms: 15.570318
...

Subtest: communication prefer-llm fallback is explicit and require-llm rejects

ok 56 - communication prefer-llm fallback is explicit and require-llm rejects

duration_ms: 20.178406
...

Subtest: project/ communication is attributed per human and agent and checked against Git evidence

ok 57 - project/ communication is attributed per human and agent and checked against Git evidence

duration_ms: 178.309318
...

Subtest: governance user-* and ai-* files become typed participant intent without ingesting ticket evidence

ok 58 - governance user-* and ai-* files become typed participant intent without ingesting ticket evidence

duration_ms: 27.681578
...

Subtest: unstructured governance participant content is rejected with an owner-specific migration warning

ok 59 - unstructured governance participant content is rejected with an owner-specific migration warning

duration_ms: 4.560874
...

Subtest: opposite wording about different explicit files is not treated as an intent conflict

ok 60 - opposite wording about different explicit files is not treated as an intent conflict

duration_ms: 7.967587
...

Subtest: missing response owners use explicit role sentinels without inventing participants

ok 61 - missing response owners use explicit role sentinels without inventing participants

duration_ms: 11.718646
...

Subtest: communication extractor reports unresolved identity instead of inventing an actor

ok 62 - communication extractor reports unresolved identity instead of inventing an actor

duration_ms: 5.699314
...

Subtest: communication extractor ignores generic generated analysis under project/

ok 63 - communication extractor ignores generic generated analysis under project/

duration_ms: 11.552685
...

Subtest: configuration converter covers JSON, TOML, Docker and CI workflow declarations

ok 64 - configuration converter covers JSON, TOML, Docker and CI workflow declarations

duration_ms: 40.551884
...

Subtest: configuration converter emits a deterministic file aggregate for an empty configuration

ok 65 - configuration converter emits a deterministic file aggregate for an empty configuration

duration_ms: 5.560812
...

Subtest: splitLines treats a trailing newline as a terminator, not an extra line

ok 66 - splitLines treats a trailing newline as a terminator, not an extra line

duration_ms: 1.689699
...

Subtest: Identical inputs produce no hunks

ok 67 - Identical inputs produce no hunks

duration_ms: 0.596966
...

Subtest: A modified line keeps both sides addressable by original line number

ok 68 - A modified line keeps both sides addressable by original line number

duration_ms: 0.403045
...

Subtest: Pure insertion and pure deletion are not reported as replacements

ok 69 - Pure insertion and pure deletion are not reported as replacements

duration_ms: 0.309081
...

Subtest: Empty-to-content and content-to-empty are handled as block changes

ok 70 - Empty-to-content and content-to-empty are handled as block changes

duration_ms: 0.306701
...

Subtest: Context width controls hunk size

ok 71 - Context width controls hunk size

duration_ms: 0.433603
...

Subtest: Nearby changes merge into a single hunk

ok 72 - Nearby changes merge into a single hunk

duration_ms: 0.329865
...

Subtest: Distant changes stay in separate hunks

ok 73 - Distant changes stay in separate hunks

duration_ms: 0.338893
...

Subtest: Oversized inputs fall back to a bounded block replace

ok 74 - Oversized inputs fall back to a bounded block replace

duration_ms: 0.617573
...

Subtest: Unified output carries a well formed hunk header

ok 75 - Unified output carries a well formed hunk header

duration_ms: 0.497057
...

Subtest: Side-by-side rows pair deletions with insertions

ok 76 - Side-by-side rows pair deletions with insertions

duration_ms: 0.328081
...

Subtest: Unbalanced change runs leave one side empty rather than misaligning

ok 77 - Unbalanced change runs leave one side empty rather than misaligning

duration_ms: 1.111211
...

Subtest: Renderers escape source markup

ok 78 - Renderers escape source markup

duration_ms: 1.490097
...

Subtest: SVG rendering caps rows and reports the remainder

ok 79 - SVG rendering caps rows and reports the remainder

duration_ms: 2.02803
...

Subtest: Reality view keys topics by target and records lane presence

ok 80 - Reality view keys topics by target and records lane presence

duration_ms: 20.325596
...

Subtest: A topic holding declared and observed records is never reported as planned-only

ok 81 - A topic holding declared and observed records is never reported as planned-only

duration_ms: 3.784482
...

Subtest: Reality coverage stays open when a shared path has unrelated capabilities

ok 82 - Reality coverage stays open when a shared path has unrelated capabilities

duration_ms: 2.264223
...

Subtest: Shared-path relations do not collapse unrelated files into one topic

ok 83 - Shared-path relations do not collapse unrelated files into one topic

duration_ms: 3.85859
...

Subtest: Reality view is deterministic for identical input

ok 84 - Reality view is deterministic for identical input

duration_ms: 3.073571
...

Subtest: Reality SVG escapes topic labels

ok 85 - Reality SVG escapes topic labels

duration_ms: 1.794578
...

Subtest: graph diff detects changed source identities, additions and SVG-safe labels

ok 86 - graph diff detects changed source identities, additions and SVG-safe labels

duration_ms: 27.5652
...

Subtest: graph diff is empty for graphs with identical evidence

ok 87 - graph diff is empty for graphs with identical evidence

duration_ms: 2.203028
...

Subtest: file diff emits deterministic unified, SVG and HTML views

ok 88 - file diff emits deterministic unified, SVG and HTML views

duration_ms: 2.315685
...

Subtest: intent-vs-reality builds an explainable SVG and Markdown projection

ok 89 - intent-vs-reality builds an explainable SVG and Markdown projection

duration_ms: 5.745638
...

Subtest: a targetless declaration is filed under the single module it links to

ok 90 - a targetless declaration is filed under the single module it links to

duration_ms: 3.413625
...

Subtest: a declaration touching several modules keeps its own topic

ok 91 - a declaration touching several modules keeps its own topic

duration_ms: 3.621179
...

Subtest: a topic about a document does not absorb the statements written in it

ok 92 - a topic about a document does not absorb the statements written in it

duration_ms: 2.96805
...

Subtest: semantically aligned configuration topics retain their evidence grade

ok 93 - semantically aligned configuration topics retain their evidence grade

duration_ms: 2.5496
...

Subtest: A record claiming line 1 is re-anchored to the line carrying its statement

ok 94 - A record claiming line 1 is re-anchored to the line carrying its statement

duration_ms: 52.353844
...

Subtest: An already correct line is kept and not reported as re-anchored

ok 95 - An already correct line is kept and not reported as re-anchored

duration_ms: 8.094266
...

Subtest: An empty target is backfilled from the statement text

ok 96 - An empty target is backfilled from the statement text

duration_ms: 12.889046
...

Subtest: A target supplied by the model is never overwritten

ok 97 - A target supplied by the model is never overwritten

duration_ms: 9.974896
...

Subtest: An unclassified action and modality are derived from the statement

ok 98 - An unclassified action and modality are derived from the statement

duration_ms: 6.824687
...

Subtest: A classified action from the model wins over the heuristic

ok 99 - A classified action from the model wins over the heuristic

duration_ms: 12.18531
...

Subtest: An action that stays unclassifiable is reported as a missing field

ok 100 - An action that stays unclassifiable is reported as a missing field

duration_ms: 7.441539
...

Subtest: A placeholder object is treated as a gap, not as content

ok 101 - A placeholder object is treated as a gap, not as content

duration_ms: 10.724431
...

Subtest: Every repair is attributable through epistemic.basis

ok 102 - Every repair is attributable through epistemic.basis

duration_ms: 11.498534
...

Subtest: deterministic documentation baseline records headings, code blocks and explicit references

ok 103 - deterministic documentation baseline records headings, code blocks and explicit references

duration_ms: 23.09708
...

Subtest: deterministic documentation preserves Polish prohibition polarity

ok 104 - deterministic documentation preserves Polish prohibition polarity

duration_ms: 9.762369
...

Subtest: deterministic documentation resolves a unique bare filename against the repository

ok 105 - deterministic documentation resolves a unique bare filename against the repository

duration_ms: 4.88998
...

Subtest: documentation prose resolves a bare filename to its repository location

ok 106 - documentation prose resolves a bare filename to its repository location

duration_ms: 4.749655
...

Subtest: a nested checkout does not shadow the repository copy of a documented file

ok 107 - a nested checkout does not shadow the repository copy of a documented file

duration_ms: 8.796265
...

Subtest: AST cache is incremental by path and source content hash

ok 108 - AST cache is incremental by path and source content hash

duration_ms: 36.346025
...

Subtest: AST cache rejects corrupt entries and recomputes authoritative records

ok 109 - AST cache rejects corrupt entries and recomputes authoritative records

duration_ms: 8.407583
...

Subtest: AST cache can be bypassed without changing extraction output

ok 110 - AST cache can be bypassed without changing extraction output

duration_ms: 5.489153
...

Subtest: successful external AST adapter is skipped on a warm manifest hit

ok 111 - successful external AST adapter is skipped on a warm manifest hit

duration_ms: 68.187217
...

Subtest: documentation chunks cache independently while provider calls remain live

ok 112 - documentation chunks cache independently while provider calls remain live

duration_ms: 35.314472
...

Subtest: generated analysis replaces its source root with a stable token

ok 113 - generated analysis replaces its source root with a stable token

duration_ms: 69.09302
...

Subtest: generated analysis root normalization refuses the filesystem root

ok 114 - generated analysis root normalization refuses the filesystem root

duration_ms: 36.309173
...

Subtest: generated analysis rejects references to untracked input

ok 115 - generated analysis rejects references to untracked input

duration_ms: 76.356725
...

Subtest: generated analysis accepts outputs independent of untracked input

ok 116 - generated analysis accepts outputs independent of untracked input

duration_ms: 90.265995
...

Subtest: generated analysis accepts an untracked filename already quoted by tracked evidence

ok 117 - generated analysis accepts an untracked filename already quoted by tracked evidence

duration_ms: 108.966923
...

Subtest: generated analysis rejects temporary paths and unavailable validators

ok 118 - generated analysis rejects temporary paths and unavailable validators

duration_ms: 82.013733
...

Subtest: generated README metadata is synchronized from package.json and stays idempotent

ok 119 - generated README metadata is synchronized from package.json and stays idempotent

duration_ms: 109.305934
...

Subtest: generated README synchronization fails closed when the template drifts

ok 120 - generated README synchronization fails closed when the template drifts

duration_ms: 36.16117
...

Subtest: generated README synchronization rejects output outside the project root

ok 121 - generated README synchronization rejects output outside the project root

duration_ms: 32.778528
...

Subtest: Git extractor emits one record per requested commit

ok 122 - Git extractor emits one record per requested commit

duration_ms: 277.011863
...

Subtest: An empty repository degrades to a warning instead of failing the run

ok 123 - An empty repository degrades to a warning instead of failing the run

duration_ms: 12.066202
...

Subtest: versioned gold dataset reports perfect offline quality and repeated-run stability

ok 124 - versioned gold dataset reports perfect offline quality and repeated-run stability

duration_ms: 177.097726
...

Subtest: gold linking reports exact-target and capability-topic quality separately

ok 125 - gold linking reports exact-target and capability-topic quality separately

duration_ms: 82.710325
...

Subtest: gold capability-topic support is large enough to detect a floor regression

ok 126 - gold capability-topic support is large enough to detect a floor regression

duration_ms: 67.985971
...

Subtest: gold known gaps are measured and kept out of precision and recall

ok 127 - gold known gaps are measured and kept out of precision and recall

duration_ms: 70.063407
...

Subtest: gold reports cross-language positives and hard negatives as a separate cohort

ok 128 - gold reports cross-language positives and hard negatives as a separate cohort

duration_ms: 73.882323
...

Subtest: gold diagnostics separate a false DONE claim from an evidenced one

ok 129 - gold diagnostics separate a false DONE claim from an evidenced one

duration_ms: 61.394025
...

Subtest: gold v1 stays evaluable after the v2 contract extension

ok 130 - gold v1 stays evaluable after the v2 contract extension

duration_ms: 27.331005
...

Subtest: gold loader rejects unsupported dataset versions

ok 131 - gold loader rejects unsupported dataset versions

duration_ms: 0.708722
...

Subtest: gold evaluator rejects unknown linking cohorts

ok 132 - gold evaluator rejects unknown linking cohorts

duration_ms: 6.01285
...

Subtest: gold v2 must declare diagnostics coverage

ok 133 - gold v2 must declare diagnostics coverage

duration_ms: 1.503149
...

Subtest: published gold schema matches the runtime contract

ok 134 - published gold schema matches the runtime contract

duration_ms: 1.821159
...

Subtest: gold evaluator rejects fixture files outside its temporary workspace

ok 135 - gold evaluator rejects fixture files outside its temporary workspace

duration_ms: 8.414502
...

Subtest: Linker connects plan, Git claim and AST fact

ok 136 - Linker connects plan, Git claim and AST fact

duration_ms: 16.304066
...

Subtest: Linker connects prose intent to a module through three grounded capability topics

ok 137 - Linker connects prose intent to a module through three grounded capability topics

duration_ms: 1.123335
...

Subtest: Linker does not connect a module on one generic topic alone

ok 138 - Linker does not connect a module on one generic topic alone

duration_ms: 1.010555
...

Subtest: An existing target path does not prove an unrelated capability

ok 139 - An existing target path does not prove an unrelated capability

duration_ms: 1.744485
...

Subtest: An existing target path plus an AST capability proves implementation

ok 140 - An existing target path plus an AST capability proves implementation

duration_ms: 1.658577
...

Subtest: Diagnostics distinguish descriptive documentation from prescriptive requirements

ok 141 - Diagnostics distinguish descriptive documentation from prescriptive requirements

duration_ms: 1.801252
...

Subtest: A changelog entry naming an extracted documentation file has release evidence

ok 142 - A changelog entry naming an extracted documentation file has release evidence

duration_ms: 1.360635
...

Subtest: Diagnostics ignore non-actionable changelog mechanics but retain release claims

ok 143 - Diagnostics ignore non-actionable changelog mechanics but retain release claims

duration_ms: 6.242388
...

Subtest: Grounded conclusion and TODO proposal contracts accept traceable values

ok 144 - Grounded conclusion and TODO proposal contracts accept traceable values

duration_ms: 7.883767
...

Subtest: Stable IDs ignore ordering noise but change with semantic content

ok 145 - Stable IDs ignore ordering noise but change with semantic content

duration_ms: 1.389677
...

Subtest: Validators reject ungrounded citations and stale semantic IDs

ok 146 - Validators reject ungrounded citations and stale semantic IDs

duration_ms: 2.044739
...

Subtest: Generation metadata exposes LLM failures instead of silently masking them

ok 147 - Generation metadata exposes LLM failures instead of silently masking them

duration_ms: 1.212347
...

Subtest: TODO proposal collections enforce dependency integrity

ok 148 - TODO proposal collections enforce dependency integrity

duration_ms: 1.255271
...

Subtest: Published JSON schemas identify all grounded output contract versions

ok 149 - Published JSON schemas identify all grounded output contract versions

duration_ms: 7.596157
...

Subtest: Blank lines and comments produce no rules

ok 150 - Blank lines and comments produce no rules

duration_ms: 1.716916
...

Subtest: A pattern without a slash matches at any depth

ok 151 - A pattern without a slash matches at any depth

duration_ms: 0.471175
...

Subtest: A leading slash anchors the pattern to the root

ok 152 - A leading slash anchors the pattern to the root

duration_ms: 0.246746
...

Subtest: A trailing slash restricts the rule to directories

ok 153 - A trailing slash restricts the rule to directories

duration_ms: 0.212838
...

Subtest: Wildcards respect path separators

ok 154 - Wildcards respect path separators

duration_ms: 0.525541
...

Subtest: Every dot-directory is excluded by .*/

ok 155 - Every dot-directory is excluded by .*/

duration_ms: 0.347759
...

Subtest: Negation re-includes a previously excluded path

ok 156 - Negation re-includes a previously excluded path

duration_ms: 0.423495
...

Subtest: Negation cannot resurrect a file inside an excluded directory

ok 157 - Negation cannot resurrect a file inside an excluded directory

duration_ms: 0.249289
...

Subtest: Last matching rule wins

ok 158 - Last matching rule wins

duration_ms: 0.352321
...

Subtest: Character classes are supported

ok 159 - Character classes are supported

duration_ms: 0.458094
...

Subtest: Paths are normalised before matching

ok 160 - Paths are normalised before matching

duration_ms: 0.338807
...

Subtest: loadIgnoreMatcher merges the three ignore files and skips missing ones

ok 161 - loadIgnoreMatcher merges the three ignore files and skips missing ones

duration_ms: 15.261195
...

Subtest: A repository without ignore files excludes nothing

ok 162 - A repository without ignore files excludes nothing

duration_ms: 1.053256
...

Subtest: The shipped .intentignore excludes build output but keeps sources

ok 163 - The shipped .intentignore excludes build output but keeps sources

duration_ms: 2.963362
...

Subtest: resolveGlobs permits one explicit .intent report without recursively scanning generated runs

ok 164 - resolveGlobs permits one explicit .intent report without recursively scanning generated runs

duration_ms: 9.655215
...

Subtest: Two unrelated AST facts sharing only a file are not linked

ok 165 - Two unrelated AST facts sharing only a file are not linked

duration_ms: 14.620301
...

Subtest: AST facts sharing a symbol are still linked despite the path rule

ok 166 - AST facts sharing a symbol are still linked despite the path rule

duration_ms: 2.20508
...

Subtest: AST details sharing only a file and generic tokens do not create a quadratic subgraph

ok 167 - AST details sharing only a file and generic tokens do not create a quadratic subgraph

duration_ms: 3.129436
...

Subtest: A file-level plan links once to the AST module aggregate instead of every detail

ok 168 - A file-level plan links once to the AST module aggregate instead of every detail

duration_ms: 7.648519
...

Subtest: A shared path still links a plan to an AST fact

ok 169 - A shared path still links a plan to an AST fact

duration_ms: 0.966149
...

Subtest: A bare filename links to a module only when its repository path is unique

ok 170 - A bare filename links to a module only when its repository path is unique

duration_ms: 1.708608
...

Subtest: A bare filename refuses ambiguous module paths

ok 171 - A bare filename refuses ambiguous module paths

duration_ms: 1.068015
...

Subtest: Relations that carry a conclusion survive alongside suppressed noise

ok 172 - Relations that carry a conclusion survive alongside suppressed noise

duration_ms: 2.265753
...

Subtest: Pair ordering stays deterministic across rebuilds

ok 173 - Pair ordering stays deterministic across rebuilds

duration_ms: 2.183748
...

Subtest: Two configuration declarations sharing only a key name are not linked

ok 174 - Two configuration declarations sharing only a key name are not linked

duration_ms: 0.533832
...

Subtest: A shared ticket still connects two configuration declarations

ok 175 - A shared ticket still connects two configuration declarations

duration_ms: 0.454136
...

Subtest: Configuration still links to documentation that describes it

ok 176 - Configuration still links to documentation that describes it

duration_ms: 0.546919
...

Subtest: Configuration file aggregate is the file-level target for an explicit documentation path

ok 177 - Configuration file aggregate is the file-level target for an explicit documentation path

duration_ms: 0.807084
...

Subtest: Configuration aggregates do not create broad capability-topic links

ok 178 - Configuration aggregates do not create broad capability-topic links

duration_ms: 0.511549
...

Subtest: a full six-stage live run passes and reports every stage

ok 179 - a full six-stage live run passes and reports every stage

duration_ms: 3.426661
...

Subtest: a stage that silently fell back to deterministic fails the check

ok 180 - a stage that silently fell back to deterministic fails the check

duration_ms: 0.517167
...

Subtest: a missing stage cannot pass as covered

ok 181 - a missing stage cannot pass as covered

duration_ms: 0.365281
...

Subtest: per-stage and total budgets are enforced separately

ok 182 - per-stage and total budgets are enforced separately

duration_ms: 0.405856
...

Subtest: live request timeout reaches the stage budget without shortening a larger override

ok 183 - live request timeout reaches the stage budget without shortening a larger override

duration_ms: 0.195128
...

Subtest: a stage reason is recorded with provider text redacted

ok 184 - a stage reason is recorded with provider text redacted

duration_ms: 0.80547
...

Subtest: history records the trend without gating on it

ok 185 - history records the trend without gating on it

duration_ms: 0.458017
...

Subtest: recorded audit history includes the current run exactly once

ok 186 - recorded audit history includes the current run exactly once

duration_ms: 0.772135
...

Subtest: history stays chronological, bounded and free of duplicate runs

ok 187 - history stays chronological, bounded and free of duplicate runs

duration_ms: 7.969106
...

Subtest: an audit converts to exactly the redacted fields history keeps

ok 188 - an audit converts to exactly the redacted fields history keeps

duration_ms: 1.174095
...

Subtest: an empty history summarizes without pretending to have measured anything

ok 189 - an empty history summarizes without pretending to have measured anything

duration_ms: 0.201234
...

Subtest: a batched run is measured per record, not per request

ok 190 - a batched run is measured per record, not per request

duration_ms: 6.317728
...

Subtest: a model whose response the validator rejected is not counted as enriched

ok 191 - a model whose response the validator rejected is not counted as enriched

duration_ms: 0.351968
...

Subtest: a failed model is a comparison result rather than a crash

ok 192 - a failed model is a comparison result rather than a crash

duration_ms: 0.89634
...

Subtest: agreement compares only records both models enriched

ok 193 - agreement compares only records both models enriched

duration_ms: 0.364976
...

Subtest: agreement is absent rather than perfect when nothing overlaps

ok 194 - agreement is absent rather than perfect when nothing overlaps

duration_ms: 0.53979
...

Subtest: the rendered comparison names the cheapest and fastest passing model

ok 195 - the rendered comparison names the cheapest and fastest passing model

duration_ms: 0.322858
...

Subtest: semantic LLM modes default to require-llm

ok 196 - semantic LLM modes default to require-llm

duration_ms: 1.889451
...

Subtest: Markdown extractor separates TODO plans and changelog claims

ok 197 - Markdown extractor separates TODO plans and changelog claims

duration_ms: 18.34059
...

Subtest: Markdown extractor preserves indented continuation lines and their source range

ok 198 - Markdown extractor preserves indented continuation lines and their source range

duration_ms: 5.822737
...

Subtest: TODO bare filenames inherit an existing directory from the heading scope

ok 199 - TODO bare filenames inherit an existing directory from the heading scope

duration_ms: 2.23172
...

Subtest: TODO resolves a bare filename only when its repository basename is unique

ok 200 - TODO resolves a bare filename only when its repository basename is unique

duration_ms: 3.861693
...

Subtest: TODO and CHANGELOG resolve the same bare filename to one repository path

ok 201 - TODO and CHANGELOG resolve the same bare filename to one repository path

duration_ms: 2.754075
...

Subtest: CHANGELOG keeps an ambiguous bare filename unresolved

ok 202 - CHANGELOG keeps an ambiguous bare filename unresolved

duration_ms: 2.601177
...

Subtest: Markdown path resolution drops paths and heading scopes outside the repository

ok 203 - Markdown path resolution drops paths and heading scopes outside the repository

duration_ms: 1.189923
...

Subtest: TODO and CHANGELOG receive audited LLM enrichment without changing structural facts

ok 204 - TODO and CHANGELOG receive audited LLM enrichment without changing structural facts

duration_ms: 25.511006
...

Subtest: Markdown enrichment corrects one rejected response and audits both attempts

ok 205 - Markdown enrichment corrects one rejected response and audits both attempts

duration_ms: 4.883962
...

Subtest: large Markdown enrichment uses bounded concurrency and keeps provider audits ordered

ok 206 - large Markdown enrichment uses bounded concurrency and keeps provider audits ordered

duration_ms: 52.81593
...

Subtest: TODO and CHANGELOG LLM fallback and require mode are explicit

ok 207 - TODO and CHANGELOG LLM fallback and require mode are explicit

duration_ms: 2.453443
...

Subtest: TODO and CHANGELOG reject structurally invalid LLM enrichments

ok 208 - TODO and CHANGELOG reject structurally invalid LLM enrichments

duration_ms: 2.612869
...

Subtest: a truncated batch is split and every record keeps its own response provenance

ok 209 - a truncated batch is split and every record keeps its own response provenance

duration_ms: 5.259736
...

Subtest: a malformed batch response splits instead of failing the whole stage

ok 210 - a malformed batch response splits instead of failing the whole stage

duration_ms: 5.600509
...

Subtest: MCP 2026 profile is stateless and exposes discovery plus complete results

ok 211 - MCP 2026 profile is stateless and exposes discovery plus complete results

duration_ms: 2.129088
...

Subtest: MCP 2026 rejects missing metadata and unsupported versions with protocol errors

ok 212 - MCP 2026 rejects missing metadata and unsupported versions with protocol errors

duration_ms: 0.790342
...

Subtest: MCP legacy profile negotiates 2025-11-25 and requires initialize

ok 213 - MCP legacy profile negotiates 2025-11-25 and requires initialize

duration_ms: 0.393414
...

Subtest: MCP exposes annotated intake command/query tools backed by the domain handler

ok 214 - MCP exposes annotated intake command/query tools backed by the domain handler

duration_ms: 12.531923
...

Subtest: An LLM record is marked as inference and keeps runtime-owned provenance

ok 215 - An LLM record is marked as inference and keeps runtime-owned provenance

duration_ms: 35.112278
...

Subtest: NL extraction corrects one rejected structured response and audits both attempts

ok 216 - NL extraction corrects one rejected structured response and audits both attempts

duration_ms: 4.343941
...

Subtest: Confidence must satisfy the provider schema instead of being silently clamped

ok 217 - Confidence must satisfy the provider schema instead of being silently clamped

duration_ms: 8.433762
...

Subtest: Source lines are clamped to the real file

ok 218 - Source lines are clamped to the real file

duration_ms: 2.602579
...

Subtest: A placeholder object is recorded as a missing field, not as content

ok 219 - A placeholder object is recorded as a missing field, not as content

duration_ms: 11.648266
...

Subtest: A real object is kept verbatim and reports no missing field

ok 220 - A real object is kept verbatim and reports no missing field

duration_ms: 2.16964
...

Subtest: The explicit unknown action is reported as a missing field

ok 221 - The explicit unknown action is reported as a missing field

duration_ms: 1.615956
...

Subtest: Both gaps are reported together

ok 222 - Both gaps are reported together

duration_ms: 1.433077
...

Subtest: Out-of-vocabulary enums are rejected instead of changing the provider intent

ok 223 - Out-of-vocabulary enums are rejected instead of changing the provider intent

duration_ms: 4.738106
...

Subtest: Rejected NL output keeps provider metadata in the failed audit

ok 224 - Rejected NL output keeps provider metadata in the failed audit

duration_ms: 2.264881
...

Subtest: The documented confidence hierarchy holds across LLM extractors

ok 225 - The documented confidence hierarchy holds across LLM extractors

duration_ms: 0.902748
...

Subtest: NL extractor produces deterministic non-LLM records

ok 226 - NL extractor produces deterministic non-LLM records

duration_ms: 10.154322
...

Subtest: NL public extraction boundary names a missing sourcePath before path resolution

ok 227 - NL public extraction boundary names a missing sourcePath before path resolution

duration_ms: 0.957022
...

Subtest: deterministic NL fallback skips Markdown headings and recognizes comparison intent

ok 228 - deterministic NL fallback skips Markdown headings and recognizes comparison intent

duration_ms: 4.197273
...

Subtest: path extraction rejects lowercase prose alternations without losing repository paths

ok 229 - path extraction rejects lowercase prose alternations without losing repository paths

duration_ms: 0.873315
...

Subtest: path extraction rejects dotted DSL fields but keeps known file extensions

ok 230 - path extraction rejects dotted DSL fields but keeps known file extensions

duration_ms: 0.263498
...

Subtest: detectModality ignores parenthetical labels and bare adjectives

ok 231 - detectModality ignores parenthetical labels and bare adjectives

duration_ms: 0.552519
...

Subtest: detectModality reads prohibitions and periphrastic obligation as requirements

ok 232 - detectModality reads prohibitions and periphrastic obligation as requirements

duration_ms: 0.562719
...

Subtest: detectPolarity does not treat without-complements as sentence negation

ok 233 - detectPolarity does not treat without-complements as sentence negation

duration_ms: 0.245307
...

Subtest: path extraction rejects HTTP routes, host paths and parent traversal

ok 234 - path extraction rejects HTTP routes, host paths and parent traversal

duration_ms: 0.297079
...

Subtest: symbol extraction rejects hostnames without losing qualified code symbols

ok 235 - symbol extraction rejects hostnames without losing qualified code symbols

duration_ms: 0.373394
...

Subtest: symbol extraction separates repository files and all-caps prose from code identifiers

ok 236 - symbol extraction separates repository files and all-caps prose from code identifiers

duration_ms: 0.210236
...

Subtest: topic keywords normalize paths, camelCase and documentation word forms

ok 237 - topic keywords normalize paths, camelCase and documentation word forms

duration_ms: 0.408341
...

Subtest: NL LLM extraction emits audited provenance and bounded DSL records

ok 238 - NL LLM extraction emits audited provenance and bounded DSL records

duration_ms: 34.182372
...

Subtest: NL LLM failure is explicit when deterministic fallback is used

ok 239 - NL LLM failure is explicit when deterministic fallback is used

duration_ms: 5.018287
...

Subtest: require-llm rejects instead of silently falling back

ok 240 - require-llm rejects instead of silently falling back

duration_ms: 0.679015
...

Subtest: OpenRouter client parses structured JSON without exposing key

ok 241 - OpenRouter client parses structured JSON without exposing key

duration_ms: 26.177383
...

Subtest: OpenRouter client preserves metadata when runtime rejects structured output

ok 242 - OpenRouter client preserves metadata when runtime rejects structured output

duration_ms: 1.736352
...

Subtest: OpenRouter client lists available models after an invalid model ID

ok 243 - OpenRouter client lists available models after an invalid model ID

duration_ms: 10.728128
...

Subtest: OpenRouter JSON timeout is not repeated as a schema fallback request

ok 244 - OpenRouter JSON timeout is not repeated as a schema fallback request

duration_ms: 0.850491
...

Subtest: OpenRouter request obeys a shared pipeline deadline without retrying

ok 245 - OpenRouter request obeys a shared pipeline deadline without retrying

duration_ms: 0.72275
...

Subtest: Documentation extractor converts OpenRouter structured output to bounded LLM records

ok 246 - Documentation extractor converts OpenRouter structured output to bounded LLM records

duration_ms: 21.878429
...

Subtest: Documentation extractor reports and enforces its chunk budget

ok 247 - Documentation extractor reports and enforces its chunk budget

duration_ms: 6.128927
...

Subtest: Documentation extractor corrects one rejected chunk and audits both responses

ok 248 - Documentation extractor corrects one rejected chunk and audits both responses

duration_ms: 5.857127
...

Subtest: Documentation extractor does not spend its correction retry on a timeout

ok 249 - Documentation extractor does not spend its correction retry on a timeout

duration_ms: 3.577088
...

Subtest: Documentation extractor exposes an audited configuration failure

ok 250 - Documentation extractor exposes an audited configuration failure

duration_ms: 0.390609
...

Subtest: Documentation extractor uses bounded concurrent OpenRouter requests

ok 251 - Documentation extractor uses bounded concurrent OpenRouter requests

duration_ms: 38.945505
...

Subtest: LLM summarizer receives graph data and preserves grounded record citations

ok 252 - LLM summarizer receives graph data and preserves grounded record citations

duration_ms: 7.219022
...

Subtest: LLM summarizer validates provider fields before creating semantic IDs

ok 253 - LLM summarizer validates provider fields before creating semantic IDs

duration_ms: 5.645427
...

Subtest: LLM summarizer diagnoses a provider that ignores the response envelope

ok 254 - LLM summarizer diagnoses a provider that ignores the response envelope

duration_ms: 3.062176
...

Subtest: LLM summarizer rejects diagnostic citations outside the supplied graph

ok 255 - LLM summarizer rejects diagnostic citations outside the supplied graph

duration_ms: 4.927535
...

Subtest: LLM summarizer prioritizes documentation over the AST payload budget

ok 256 - LLM summarizer prioritizes documentation over the AST payload budget

duration_ms: 132.334546
...

Subtest: deterministic summary presents AST module aggregates instead of low-level calls

ok 257 - deterministic summary presents AST module aggregates instead of low-level calls

duration_ms: 2.850398
...

Subtest: The summarizer grounds a fabricated record citation from its diagnostic

ok 258 - The summarizer grounds a fabricated record citation from its diagnostic

duration_ms: 4.055261
...

Subtest: The summarizer still fails when the retry fabricates a diagnostic again

ok 259 - The summarizer still fails when the retry fabricates a diagnostic again

duration_ms: 3.300453
...

Subtest: variable contracts and operation plans have deterministic content-bound IDs

ok 260 - variable contracts and operation plans have deterministic content-bound IDs

duration_ms: 5.614727
...

Subtest: every variable grants Founder read/write authority and immutable variables reject other writers

ok 261 - every variable grants Founder read/write authority and immutable variables reject other writers

duration_ms: 0.720905
...

Subtest: plans reject undeclared parameters, actor visibility gaps and payload secrets

ok 262 - plans reject undeclared parameters, actor visibility gaps and payload secrets

duration_ms: 1.363876
...

Subtest: safety-sensitive commands require a Founder decision, a human boundary and verification

ok 263 - safety-sensitive commands require a Founder decision, a human boundary and verification

duration_ms: 1.057251
...

Subtest: plan hash detects semantic tampering

ok 264 - plan hash detects semantic tampering

duration_ms: 1.046037
...

Subtest: compiler emits the exact governed envelope without an execution surface

ok 265 - compiler emits the exact governed envelope without an execution surface

duration_ms: 1.065453
...

Subtest: runtime draft boundaries ignore lifecycle and identity fields injected by untyped callers

ok 266 - runtime draft boundaries ignore lifecycle and identity fields injected by untyped callers

duration_ms: 0.609059
...

Subtest: compiler fails closed on extra, stale, wrong-source and wrong-type bindings

ok 267 - compiler fails closed on extra, stale, wrong-source and wrong-type bindings

duration_ms: 1.471806
...

Subtest: file boundary writes one private envelope atomically and refuses overwrite

ok 268 - file boundary writes one private envelope atomically and refuses overwrite

duration_ms: 13.57971
...

Subtest: Offline pipeline writes a complete run

ok 269 - Offline pipeline writes a complete run

duration_ms: 141.186458
...

Subtest: Pipeline persists synthesis, validation and review patch, then registers approval receipt

ok 270 - Pipeline persists synthesis, validation and review patch, then registers approval receipt

duration_ms: 30.003963
...

Subtest: Pipeline integrates multi-participant communication into graph, diagnostics, reality and run artifacts

ok 271 - Pipeline integrates multi-participant communication into graph, diagnostics, reality and run artifacts

duration_ms: 35.716437
...

Subtest: Pipeline require-llm task synthesis failure is audited and never publishes latest

ok 272 - Pipeline require-llm task synthesis failure is audited and never publishes latest

duration_ms: 12.271371
...

Subtest: Pipeline persists an audited failure when communication require-llm cannot run

ok 273 - Pipeline persists an audited failure when communication require-llm cannot run

duration_ms: 12.266632
...

Subtest: Pipeline persists communication stage failure and does not publish latest

ok 274 - Pipeline persists communication stage failure and does not publish latest

duration_ms: 20.504663
...

Subtest: Pipeline persists a failed manifest when NL require-llm aborts

ok 275 - Pipeline persists a failed manifest when NL require-llm aborts

duration_ms: 1.991047
...

Subtest: Pipeline persists a failed manifest when Markdown require-llm aborts

ok 276 - Pipeline persists a failed manifest when Markdown require-llm aborts

duration_ms: 8.840128
...

Subtest: Pipeline persists a failed manifest for an unexpected summary failure

ok 277 - Pipeline persists a failed manifest for an unexpected summary failure

duration_ms: 10.275784
...

Subtest: Proposal validation reports existing TODO duplicates and orders dependencies before priority

ok 278 - Proposal validation reports existing TODO duplicates and orders dependencies before priority

duration_ms: 17.973545
...

Subtest: Proposal validation rejects dependency cycles and whitespace-only criterion duplicates

ok 279 - Proposal validation rejects dependency cycles and whitespace-only criterion duplicates

duration_ms: 2.345688
...

Subtest: Python package executes the local TypeScript reality runtime without a server

ok 280 - Python package executes the local TypeScript reality runtime without a server

duration_ms: 2119.643435
...

Subtest: runtime cycle extractor turns probes, violations, drift and proposals into records

ok 281 - runtime cycle extractor turns probes, violations, drift and proposals into records

duration_ms: 18.1615
...

Subtest: measurements enter as fact and the cycle own suggestions as inference

ok 282 - measurements enter as fact and the cycle own suggestions as inference

duration_ms: 5.913802
...

Subtest: a violated expectation is negative and blocked, a healthy probe positive and verified

ok 283 - a violated expectation is negative and blocked, a healthy probe positive and verified

duration_ms: 2.619289
...

Subtest: a probe that could not run is unknown, not failing

ok 284 - a probe that could not run is unknown, not failing

duration_ms: 1.520496
...

Subtest: a watched path lets a runtime fact link to the code that produces it

ok 285 - a watched path lets a runtime fact link to the code that produces it

duration_ms: 17.560164
...

Subtest: a document that is not an autonom cycle is refused by schema, not silently empty

ok 286 - a document that is not an autonom cycle is refused by schema, not silently empty

duration_ms: 2.44982
...

Subtest: Runtime validator enforces the complete Intent DSL enum and object contract

ok 287 - Runtime validator enforces the complete Intent DSL enum and object contract

duration_ms: 4.338098
...

Subtest: Linker and remote action boundary reject malformed records before graph construction

ok 288 - Linker and remote action boundary reject malformed records before graph construction

duration_ms: 5.811759
...

Subtest: Graph validator rejects invalid relations and inconsistent statistics

ok 289 - Graph validator rejects invalid relations and inconsistent statistics

duration_ms: 2.519089
...

[t2c:a2a] listening on 127.0.0.1:36077

Subtest: diff UI and TypeScript/Python SDKs use the live backend runtime

ok 290 - diff UI and TypeScript/Python SDKs use the live backend runtime

duration_ms: 234.747876
...

Subtest: MCP/A2A action boundary rejects traversal and symlink escapes

ok 291 - MCP/A2A action boundary rejects traversal and symlink escapes

duration_ms: 24.973522
...

Subtest: bounded retrieval cannot create a relation until a grounded reranker accepts it

ok 292 - bounded retrieval cannot create a relation until a grounded reranker accepts it

duration_ms: 18.962655
...

Subtest: reranker fails closed on ungrounded quotes and more than one accepted module

ok 293 - reranker fails closed on ungrounded quotes and more than one accepted module

duration_ms: 4.074137
...

Subtest: OpenRouter reranking is required, structured and reusable only through an identity-bound cache

ok 294 - OpenRouter reranking is required, structured and reusable only through an identity-bound cache

duration_ms: 74.179342
...

Subtest: published semantic reranker schemas expose the versioned bounded contracts

ok 295 - published semantic reranker schemas expose the versioned bounded contracts

duration_ms: 1.713597
...

Subtest: provider response validation diagnoses the exact property without coercion

ok 296 - provider response validation diagnoses the exact property without coercion

duration_ms: 0.609914
...

Subtest: one structured contract emits the provider schema and parses the same value

ok 297 - one structured contract emits the provider schema and parses the same value

duration_ms: 2.484289
...

Subtest: structured parsing fails closed with the exact response path

ok 298 - structured parsing fails closed with the exact response path

duration_ms: 0.874779
...

Subtest: object uniqueness uses canonical JSON identity rather than property order

ok 299 - object uniqueness uses canonical JSON identity rather than property order

duration_ms: 0.348751
...

Subtest: a short NL symbol resolves to its only AST owner

ok 300 - a short NL symbol resolves to its only AST owner

duration_ms: 20.635731
...

Subtest: an ambiguous short NL symbol does not pretend that either AST owner is selected

ok 301 - an ambiguous short NL symbol does not pretend that either AST owner is selected

duration_ms: 3.748566
...

Subtest: an explicit path selects one owner of an otherwise ambiguous symbol

ok 302 - an explicit path selects one owner of an otherwise ambiguous symbol

duration_ms: 2.217069
...

Subtest: a qualified symbol selects its exact AST declaration without a path

ok 303 - a qualified symbol selects its exact AST declaration without a path

duration_ms: 1.081319
...

Subtest: a symbol and explicit path conflict reports the observed AST location

ok 304 - a symbol and explicit path conflict reports the observed AST location

duration_ms: 0.977542
...

Subtest: missingFields diagnostics prescribe a concrete edit for every known gap

ok 305 - missingFields diagnostics prescribe a concrete edit for every known gap

duration_ms: 0.724937
...

Subtest: Target normalization canonicalizes paths, symbols and cross-language separators

ok 306 - Target normalization canonicalizes paths, symbols and cross-language separators

duration_ms: 2.138032
...

Subtest: Qualified AST symbols align with short plan and documentation targets

ok 307 - Qualified AST symbols align with short plan and documentation targets

duration_ms: 17.368799
...

Subtest: Structured task synthesis materializes stable, grounded contracts with a complete audit

ok 308 - Structured task synthesis materializes stable, grounded contracts with a complete audit

duration_ms: 45.99265
...

Subtest: blank response-local proposal keys are rejected instead of invented by the runtime

ok 309 - blank response-local proposal keys are rejected instead of invented by the runtime

duration_ms: 6.79877
...

Subtest: prefer-llm exposes raw diagnostic actions without claiming semantic task generation

ok 310 - prefer-llm exposes raw diagnostic actions without claiming semantic task generation

duration_ms: 1.267921
...

Subtest: communication divergence is grounded in task synthesis without treating agent claims as facts

ok 311 - communication divergence is grounded in task synthesis without treating agent claims as facts

duration_ms: 5.209256
...

Subtest: task synthesis defaults to require-llm and fails explicitly without a provider

ok 312 - task synthesis defaults to require-llm and fails explicitly without a provider

duration_ms: 1.273814
...

Subtest: invalid structured LLM citations are rejected or visibly degraded according to mode

ok 313 - invalid structured LLM citations are rejected or visibly degraded according to mode

duration_ms: 10.452399
...

Subtest: task synthesis timeout is audited and never retried as a format fallback

ok 314 - task synthesis timeout is audited and never retried as a format fallback

duration_ms: 16.464353
...

Subtest: A fabricated record citation is grounded from its cited diagnostic without a retry

ok 315 - A fabricated record citation is grounded from its cited diagnostic without a retry

duration_ms: 4.844261
...

Subtest: A fabricated diagnostic still fails after the corrective retry

ok 316 - A fabricated diagnostic still fails after the corrective retry

duration_ms: 5.583099
...

Subtest: TensorFlow remains an explicit fallback when the isolated adapter is not installed

ok 317 - TensorFlow remains an explicit fallback when the isolated adapter is not installed

duration_ms: 7.394806
...

Subtest: TODO patch rendering is stable, dependency-first and excludes classified duplicates

ok 318 - TODO patch rendering is stable, dependency-first and excludes classified duplicates

duration_ms: 19.761122
...

Subtest: empty and duplicate-only results render an explicit no-op patch

ok 319 - empty and duplicate-only results render an explicit no-op patch

duration_ms: 3.543215
...

Subtest: apply rejects missing or wrong approval, stale TODO and a tampered patch

ok 320 - apply rejects missing or wrong approval, stale TODO and a tampered patch

duration_ms: 16.394695
...

Subtest: approved apply is atomic, receipt-backed and idempotent

ok 321 - approved apply is atomic, receipt-backed and idempotent

duration_ms: 9.500674
...

Subtest: service actions execute LLM propose -> render -> approved apply with scoped artifacts

ok 322 - service actions execute LLM propose -> render -> approved apply with scoped artifacts

duration_ms: 38.464243
...

Subtest: scanTree prunes ignored directories and records file signatures

ok 323 - scanTree prunes ignored directories and records file signatures

duration_ms: 16.387029
...

Subtest: diffSnapshots classifies additions, modifications and removals

ok 324 - diffSnapshots classifies additions, modifications and removals

duration_ms: 0.443609
...

Subtest: describeDelta truncates long change lists

ok 325 - describeDelta truncates long change lists

duration_ms: 0.200925
...

Subtest: An unchanged tree produces exactly one report and then stays quiet

ok 326 - An unchanged tree produces exactly one report and then stays quiet

duration_ms: 3.76991
...

Subtest: Reports are rate limited to one per interval no matter how often files change

ok 327 - Reports are rate limited to one per interval no matter how often files change

duration_ms: 26.688159
...

Subtest: A change is reported once the interval has elapsed

ok 328 - A change is reported once the interval has elapsed

duration_ms: 2.817886
...

Subtest: Ignored files never trigger a report

ok 329 - Ignored files never trigger a report

duration_ms: 4.678022
...

Subtest: A failing report is surfaced and does not stop the watcher

ok 330 - A failing report is surfaced and does not stop the watcher

duration_ms: 1.51215
...

Subtest: --no-initial-report waits for a real change

ok 331 - --no-initial-report waits for a real change

duration_ms: 1.251563
...

Subtest: Communication changes trigger watch and coalesce under the existing report rate limit

ok 332 - Communication changes trigger watch and coalesce under the existing report rate limit

duration_ms: 3.311932
...

Subtest: workflow verifier rejects duplicate top-level YAML keys

ok 333 - workflow verifier rejects duplicate top-level YAML keys

duration_ms: 77.272234
...

Subtest: workspace headline trend ignores AST-only topic and source churn

ok 334 - workspace headline trend ignores AST-only topic and source churn

duration_ms: 1.456164
...

Subtest: workspace comparison measures origin/main against uncommitted filesystem intent

ok 335 - workspace comparison measures origin/main against uncommitted filesystem intent

duration_ms: 326.429539
...
1..335

tests 335

suites 0

pass 334

fail 0

cancelled 0

skipped 1

todo 0

duration_ms 10600.724353: 335 tests, 334 pass, 0 fail, 1 local JDK skip

• fresh Docker : pass
• Docker : independently owned stale , expected exit 101 at

Follow-up failure-path validation

• actual LiteLLM HTTP stack + local 404 endpoint: NotFoundError in ~705 ms, exactly one request, max_tokens=8192
• slow local endpoint + 0.5 s request limit: Timeout in ~799 ms, exactly one request
• fresh npm run verify: 335 tests, 334 pass, 0 fail, 1 local JDK skip
• fresh Docker e2e-core: pass
• Docker e2e-full: independently owned stale sdk/rust/Cargo.lock, expected exit 101 at cargo fetch --locked