{ "version": 1, "id": "Circe-accepted-turn", "repo": "/home/abstergo/Programming/Circe", "repoName": "Circe", "scope": "accepted-turn", "title": "Circe PR #2 review, round 6: follow-up ready", "sinceCommit": "3d3bf9bb3403a484eeeb711512607d3b8fe76053", "lastReviewCommit": "672b86d991b1f23d6a7dc5fc4be38542ce59a0e7", "reviewRound": 6, "loc": { "added": 5261, "removed": 158 }, "updatedAt": "2026-09-22T21:14:18.136742+00:00", "htmlPath": "/home/abstergo/Documents/reviews/Circe-accepted-turn-abbyview.html", "jsonPath": "/home/abstergo/Documents/reviews/Circe-accepted-turn-abbyview.json", "verdict": "APPROVED", "confidence": { "recommendation": "PR #7 is ready to merge for the reviewed scope", "rationale": "F44-F45 verified fixed, no new findings in the incremental change. 279 repo tests, typecheck and all 59 independent probes pass. No numeric score, as requested." }, "actionableOpen": 0, "currentBlockingFindings": 0, "currentAdditionalFindings": 0, "deferredOpen": 0, "baseline": { "frozenAt": "2026-09-22T12:41:53.536939+00:00", "title": "Circe PR #2 accepted-turn review", "lede": "Good architectural direction, but the compiler still authorizes actions outside its claimed contract.", "walkthrough": { "prose": "The PR adds an accepted-turn bridge between workflow judgments and typed kernel proposals.", "bullets": [ "semanticTurnFromRun maps accepted spec ids to trace ids and carries run/frame identity into the turn.", "compileSemanticTurn checks citation provenance and capability selection, computes exclusions, derives observed versions and suppresses identical commands.", "The latest refactor constructs commands directly, leaving act/operation checks only in propositionCommand.", "The compiler now recognizes argument refs, but the kernel still stores read evidence only on slot bindings.", "The new evals demonstrate contract behavior with scripted decisions and manually constructed propositions; real interpretation remains a later integration concern." ] }, "filesChanged": [ { "path": "packages/context/src/types.ts", "added": 2, "removed": 0, "delta": "+2 / -0" }, { "path": "packages/evals/src/accepted-turn-adversarial.test.ts", "added": 350, "removed": 0, "delta": "+350 / -0" }, { "path": "packages/evals/src/accepted-turn-pipeline.test.ts", "added": 542, "removed": 0, "delta": "+542 / -0" }, { "path": "packages/perception/src/proposals.ts", "added": 631, "removed": 84, "delta": "+631 / -84" }, { "path": "packages/perception/src/trace.ts", "added": 1, "removed": 0, "delta": "+1 / -0" }, { "path": "packages/perception/src/types.ts", "added": 5, "removed": 0, "delta": "+5 / -0" } ], "flowDiagram": "Workflow run \u2192 accepted decision ids + traces \u2192 whole-turn validation \u2192 evidence-bound commands \u2192 kernel validation/freeze \u2192 executor", "sinceCommit": "3d3bf9bb3403a484eeeb711512607d3b8fe76053", "loc": { "added": 1531, "removed": 84 } }, "roundDeltas": [ { "round": 2, "at": "2026-09-22T12:59:04.708674+00:00", "commit": "34685ba2cbb50d7d029437a040d1d3a6dfe0eac6", "filesChanged": [ { "path": "packages/core/src/runtime.ts", "added": 27, "removed": 3, "delta": "+27 / -3" }, { "path": "packages/core/src/types.ts", "added": 6, "removed": 0, "delta": "+6 / -0" }, { "path": "packages/evals/src/accepted-turn-pipeline.test.ts", "added": 1, "removed": 1, "delta": "+1 / -1" }, { "path": "packages/evals/src/accepted-turn-review2.test.ts", "added": 263, "removed": 0, "delta": "+263 / -0" }, { "path": "packages/perception/src/proposals.ts", "added": 114, "removed": 32, "delta": "+114 / -32" } ], "loc": { "added": 411, "removed": 36 }, "prose": "Verified all original F31-F37 cases. The added shared act and shape checks work, and argument evidence now reaches validation. One dependency path remains outside whole-turn rejection tracking, and the evidence copy introduces revision and target-effect regressions.", "bullets": [ "234 PR #2 tests and typecheck pass.", "All 41 independent earlier reproductions pass.", "Four new required-behavior cases fail; the genuine shared-dependency control passes.", "All three evidence regression cases pass at 91c123b.", "Stack tip 623e2d7 passes 281 tests and typecheck, but also reproduces all four new failures." ] }, { "round": 3, "at": "2026-09-22T17:13:11.179624+00:00", "commit": "2cb82c8582ff2edf68fc8f47a5ec90d2aa1fada8", "filesChanged": [ { "path": "packages/core/src/runtime.ts", "added": 72, "removed": 7, "delta": "+72 / -7" }, { "path": "packages/evals/src/accepted-turn-review2-round2.test.ts", "added": 258, "removed": 0, "delta": "+258 / -0" }, { "path": "packages/perception/src/proposals.ts", "added": 214, "removed": 209, "delta": "+214 / -209" } ], "loc": { "added": 544, "removed": 216 }, "prose": "The shared evaluateTurn resolves the demonstrated dependency entry-point drift. Original evidence revision and independent-target cases pass. New overlap and revision tests expose lost required dependencies.", "bullets": [ "239 PR #2 tests and typecheck pass.", "All 46 independent prior reproductions pass.", "Three new cases fail at 2cb82c8 and pass at 34685ba.", "Stack tip 9d78c7e passes 286 tests and typecheck, but reproduces all three new failures." ] }, { "round": 4, "at": "2026-09-22T20:51:05.040219+00:00", "commit": "a5d74b5790867568948c579ec07e2756fd6292fc", "filesChanged": [ { "path": "packages/context/src/assembly.ts", "added": 1, "removed": 0, "delta": "+1 / -0" }, { "path": "packages/context/src/overlay.ts", "added": 7, "removed": 0, "delta": "+7 / -0" }, { "path": "packages/context/src/types.ts", "added": 6, "removed": 0, "delta": "+6 / -0" }, { "path": "packages/core/src/runtime.ts", "added": 502, "removed": 92, "delta": "+502 / -92" }, { "path": "packages/core/src/types.ts", "added": 26, "removed": 3, "delta": "+26 / -3" }, { "path": "packages/evals/src/accepted-turn-evidence-ownership.test.ts", "added": 1869, "removed": 0, "delta": "+1869 / -0" }, { "path": "packages/evals/src/accepted-turn-review2-round3.test.ts", "added": 243, "removed": 0, "delta": "+243 / -0" }, { "path": "packages/perception/src/proposals.ts", "added": 88, "removed": 4, "delta": "+88 / -4" }, { "path": "packages/resolvers-deterministic/src/classify.ts", "added": 26, "removed": 8, "delta": "+26 / -8" } ], "loc": { "added": 2768, "removed": 107 }, "prose": "Reviewed the 16 self-review commits after 2cb82c8 for correctness and unnecessary complexity. F40-F41 are fixed. Added positive controls for unrelated completed intentions and descriptor/overlay agreement; they identify two new regressions.", "bullets": [ "PR #2 is already merged as f823054. Reviewed final head a5d74b5.", "274 baseline tests and typecheck pass.", "All 49 prior independent probes pass.", "Three new valid-flow cases fail, two positive controls pass. All five pass at 2cb82c8.", "Current stack tip 8cea673 reproduces the same three failures. Later PRs were not fully reviewed." ] }, { "round": 5, "at": "2026-09-22T21:05:35.964356+00:00", "commit": "ae4ae3de85905bc2f7c2752679168991cf51561c", "filesChanged": [ { "path": "packages/core/src/runtime.ts", "added": 32, "removed": 20, "delta": "+32 / -20" }, { "path": "packages/core/src/types.ts", "added": 25, "removed": 1, "delta": "+25 / -1" }, { "path": "packages/evals/src/accepted-turn-evidence-ownership.test.ts", "added": 107, "removed": 2, "delta": "+107 / -2" }, { "path": "packages/perception/src/proposals.ts", "added": 30, "removed": 7, "delta": "+30 / -7" }, { "path": "packages/perception/src/types.ts", "added": 2, "removed": 0, "delta": "+2 / -0" } ], "loc": { "added": 196, "removed": 30 }, "prose": "Reviewed PR #7 follow-up to merged PR #2. Original F42-F43 cases now pass. The explicit continuation identity is dropped by the semantic-frame bridge and omitted from the canonical deduplication key.", "bullets": [ "277 repo tests and typecheck pass.", "All 54 earlier independent cases pass unchanged.", "Two new identity cases fail; a named-current positive control passes.", "Fix scope remains narrow; no architectural rewrite recommended." ] }, { "round": 6, "at": "2026-09-22T21:14:18.136742+00:00", "commit": "672b86d991b1f23d6a7dc5fc4be38542ce59a0e7", "filesChanged": [ { "path": "packages/core/src/runtime.ts", "added": 9, "removed": 0, "delta": "+9 / -0" }, { "path": "packages/evals/src/accepted-turn-evidence-ownership.test.ts", "added": 116, "removed": 0, "delta": "+116 / -0" }, { "path": "packages/perception/src/proposals.ts", "added": 1, "removed": 0, "delta": "+1 / -0" } ], "loc": { "added": 126, "removed": 0 }, "prose": "The final follow-up carries expectedIntentId through semantic ingestion, rejects malformed identities instead of dropping them, and includes identity in command deduplication. Original failures and two additional boundary controls pass.", "bullets": [ "279 repo tests and typecheck pass.", "All 57 earlier probes plus two additional boundary controls pass, 59 total.", "Ten production lines added, plus 116 regression-test lines. No new abstraction or broader rewrite.", "No new review findings; later stack PRs remain outside this approval." ] } ], "findings": [ { "id": "f31", "status": "fixed", "fixedAt": "34685ba2cbb50d7d029437a040d1d3a6dfe0eac6", "reviewRound": 1, "title": "The whole-turn compiler turns queries into mutations", "file": "packages/perception/src/proposals.ts", "lines": "605-616", "severity": "major", "pBadge": "P1", "type": "logic", "section": "actionable", "category": "correctness", "bodyParagraphs": [ "compileSemanticTurn no longer calls propositionCommand. It calls buildCommand directly after citation and relation checks, but never checks that the act is request or that the operation is reference or slot-fill. Those checks exist only in the other API at lines 445-449. The latest 91c123b refactor removed their effective use by the whole-turn compiler.", "A resolved query/inspect proposition with an accepted task.resume capability citation and an observed target compiles to a propose command and reaches the executor. The same happens for query/reference and request/compose. All three required-behavior tests fail on an actual invocation, not merely on command construction.", "The existing query regression supplies a referent citation rather than the capability citation needed to get this far, so it passes for a different rejection reason. The kernel receives no dialogue act and cannot repair the lost distinction." ], "fix": "Use one executable-proposition validator for both entry points, including capability presence and act/operation eligibility. Test non-executable acts with otherwise valid citations and evidence.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/perception/src/proposals.ts` around lines 605-616, compileSemanticTurn no longer calls propositionCommand. It calls buildCommand directly after citation and relation checks, but never checks that the act is request or that the operation is reference or slot-fill. Those checks exist only in the other API at lines 445-449. The latest 91c123b refactor removed their effective use by the whole-turn compiler. Use one executable-proposition validator for both entry points, including capability presence and act/operation eligibility. Test non-executable acts with otherwise valid citations and evidence. Reproduce R1 using `/tmp/circe-pr2-evidence/probes.test.ts` at PR #2 head 91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893. Keep valid requests working and run the full suite and typecheck after the change.", "followUp": false, "source": "605: const issue = citationIssue(proposition, turn, acceptedDecisions, tracesById, tracesAvailable);\n606: if (issue) {\n607: rejected.push({ index, reason: issue });\n608: return;\n609: }\n610: const relation = relationIssue(turn, proposition, index, accepted);\n611: if (relation) {\n612: rejected.push({ index, reason: relation });\n613: return;\n614: }\n615: \n616: const declared = slotsOf(proposition);", "test": "R1", "commit": "91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893", "resolutionNote": "The original reproduction passes independently at 34685ba. Adjacent remaining paths and evidence regressions are recorded separately as F38-F39, without rewriting this finding." }, { "id": "f32", "status": "fixed", "fixedAt": "34685ba2cbb50d7d029437a040d1d3a6dfe0eac6", "reviewRound": 1, "title": "Argument reference evidence disappears at the kernel boundary", "file": "packages/perception/src/proposals.ts", "lines": "665-676", "severity": "major", "pBadge": "P1", "type": "logic", "section": "actionable", "category": "data-integrity", "bodyParagraphs": [ "The new argument-reference path adds versions for argument-only refs to command.expectedVersions. However, the kernel in runtime.ts:1051-1057 filters that list to refs present in named slot bindings. Its commit path validates and freezes only binding expectations. The new compiler therefore emits evidence that its actual consumer silently drops.", "The reproduction compiles slot target A plus arguments.secondary = B. The command correctly includes versions for A and B, but authority validation receives only A and the executor receives the B argument without a B precondition. A second test makes authority validation return stale for B if asked. Execution still happens because B is never checked.", "This breaks the newly claimed coverage of every executable argument reference. Seeing B in the frame at compile time does not establish that B is current at dispatch. No real backend or persistence adapter is required to reproduce the gap." ], "fix": "Until command-wide read preconditions survive validation and freezing, reject argument refs outside executable bindings. Alternatively add explicit command-level read dependencies and preserve them through every effect unit; do not silently invent target slots that change execution semantics.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/perception/src/proposals.ts` around lines 665-676, The new argument-reference path adds versions for argument-only refs to command.expectedVersions. However, the kernel in runtime.ts:1051-1057 filters that list to refs present in named slot bindings. Its commit path validates and freezes only binding expectations. The new compiler therefore emits evidence that its actual consumer silently drops. Until command-wide read preconditions survive validation and freezing, reject argument refs outside executable bindings. Alternatively add explicit command-level read dependencies and preserve them through every effect unit; do not silently invent target slots that change execution semantics. Reproduce R2 / R2b using `/tmp/circe-pr2-evidence/probes.test.ts` at PR #2 head 91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893. Keep valid requests working and run the full suite and typecheck after the change.", "followUp": false, "source": "665: // Every executable ref must carry the version the interpretation read.\n666: const evidence = expectedVersionsFromFrame(options.frame, [\n667: ...constrained.flatMap((slot) => slot.refs),\n668: ...argumentRefs,\n669: ]);\n670: for (const ref of [...constrained.flatMap((slot) => slot.refs), ...argumentRefs]) {\n671: if (!evidence.some((precondition) => entityRefKey(precondition.ref) === entityRefKey(ref))) {\n672: rejected.push({ index, reason: `no frame observation for ${entityRefKey(ref)}` });\n673: return;\n674: }\n675: }\n676: const command = buildCommand({ proposition, commandId, slots: constrained, argumentRefs, evidence });", "test": "R2 / R2b", "commit": "91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893", "resolutionNote": "The original reproduction passes independently at 34685ba. Adjacent remaining paths and evidence regressions are recorded separately as F38-F39, without rewriting this finding." }, { "id": "f33", "status": "fixed", "fixedAt": "34685ba2cbb50d7d029437a040d1d3a6dfe0eac6", "reviewRound": 1, "title": "Rejected prerequisites do not block dependent commands", "file": "packages/perception/src/proposals.ts", "lines": "355-359", "severity": "major", "pBadge": "P1", "type": "logic", "section": "actionable", "category": "correctness", "bodyParagraphs": [ "relationIssue treats a dependsOn target as satisfied when its cited decisions are accepted. It never checks whether that proposition survives evidence validation, prohibitions or other compilation checks. compileSemanticTurn then returns the dependent command with no representation of the relation.", "The reproduction has proposition 0 request B without a frame observation, and proposition 1 request A with dependsOn:0. The compiler explicitly rejects proposition 0 for missing evidence yet emits proposition 1. A caller executing the returned commands cannot discover the missing prerequisite from that command.", "Acceptance of a judgment is weaker than acceptance of the proposition it supports. A whole-turn compiler needs to preserve that distinction. This finding concerns a rejected prerequisite at compilation; it does not require adding a general workflow scheduler to this PR." ], "fix": "Compute proposition validity before emitting commands and propagate rejection through dependency edges. Reject executable dependency forms that the command contract cannot preserve; define semantic-only dependencies explicitly rather than silently treating rejected requests as satisfied.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/perception/src/proposals.ts` around lines 355-359, relationIssue treats a dependsOn target as satisfied when its cited decisions are accepted. It never checks whether that proposition survives evidence validation, prohibitions or other compilation checks. compileSemanticTurn then returns the dependent command with no representation of the relation. Compute proposition validity before emitting commands and propagate rejection through dependency edges. Reject executable dependency forms that the command contract cannot preserve; define semantic-only dependencies explicitly rather than silently treating rejected requests as satisfied. Reproduce R3 using `/tmp/circe-pr2-evidence/probes.test.ts` at PR #2 head 91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893. Keep valid requests working and run the full suite and typecheck after the change.", "followUp": false, "source": "355: const related = turn.propositions[relation.targetIndex];\n356: if (!related) return \"malformed proposition: relation target is missing\";\n357: if (relation.kind === \"dependsOn\") {\n358: if (!isAccepted(related)) return \"a dependency proposition was not accepted\";\n359: }", "test": "R3", "commit": "91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893", "resolutionNote": "The original reproduction passes independently at 34685ba. Adjacent remaining paths and evidence regressions are recorded separately as F38-F39, without rewriting this finding." }, { "id": "f34", "status": "fixed", "fixedAt": "34685ba2cbb50d7d029437a040d1d3a6dfe0eac6", "reviewRound": 1, "title": "A dangling prohibition is silently ignored", "file": "packages/perception/src/proposals.ts", "lines": "123-135", "severity": "major", "pBadge": "P1", "type": "logic", "section": "actionable", "category": "correctness", "bodyParagraphs": [ "shapeIssue checks that a relation targetIndex is an integer, but does not check its bounds. Bounds are checked later in relationIssue, after exclusions have already returned from the compilation loop. The whole-turn malformed-input rule consequently misses dangling relations, including prohibition-only propositions.", "The reproduction submits an exclusion with no direct refs and an excludes relation targeting index 99 in a two-proposition turn. The exclusion is labeled an exclusion constraint, while the other proposition compiles into a command. The malformed prohibition was never validated or rejected as malformed.", "The PR expressly promises that malformed relations make the whole turn uncompilable because ignoring them can discard intended prohibitions. The current test covers a string index, but an out-of-range integer passes that shape gate." ], "fix": "Validate relation bounds and target existence for the entire turn before analyzing exclusions or constructing any commands. Apply the same validation to exclusion-only propositions and both public entry points.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/perception/src/proposals.ts` around lines 123-135, shapeIssue checks that a relation targetIndex is an integer, but does not check its bounds. Bounds are checked later in relationIssue, after exclusions have already returned from the compilation loop. The whole-turn malformed-input rule consequently misses dangling relations, including prohibition-only propositions. Validate relation bounds and target existence for the entire turn before analyzing exclusions or constructing any commands. Apply the same validation to exclusion-only propositions and both public entry points. Reproduce R4 using `/tmp/circe-pr2-evidence/probes.test.ts` at PR #2 head 91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893. Keep valid requests working and run the full suite and typecheck after the change.", "followUp": false, "source": "123: if (proposition.relations !== undefined) {\n124: if (!Array.isArray(proposition.relations)) {\n125: return \"malformed proposition: relations is not an array\";\n126: }\n127: for (const relation of proposition.relations) {\n128: if (\n129: !isRecord(relation) ||\n130: typeof relation.kind !== \"string\" ||\n131: !RELATION_KINDS.has(relation.kind) ||\n132: !Number.isInteger(relation.targetIndex)\n133: ) {\n134: return \"malformed proposition: relation is not a typed, integer-indexed reference\";\n135: }", "test": "R4", "commit": "91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893", "resolutionNote": "The original reproduction passes independently at 34685ba. Adjacent remaining paths and evidence regressions are recorded separately as F38-F39, without rewriting this finding." }, { "id": "f35", "status": "fixed", "fixedAt": "34685ba2cbb50d7d029437a040d1d3a6dfe0eac6", "reviewRound": 1, "title": "The explicit proposition index can bypass prohibitions", "file": "packages/perception/src/proposals.ts", "lines": "453-469", "severity": "major", "pBadge": "P1", "type": "logic", "section": "actionable", "category": "correctness", "bodyParagraphs": [ "The new index option is trusted without checking its range or confirming that turn.propositions[index] matches the supplied proposition. The code uses that index to look up turn-wide prohibitions, while it constructs the command from the separate proposition argument.", "The reproduction places a relation-only exclusion at index 0 prohibiting the request at index 1. Calling propositionCommand with that request and index:0 returns an executable command. Checking the correct position would reject it. Rehydration or filtering code can therefore make the guard inspect one proposition while executing another.", "This is a defect in the new exported low-level API even though compileSemanticTurn no longer calls it. The API explicitly exists for callers holding accepted decisions and advertises the same turn-wide prohibitions." ], "fix": "Require an integer index within bounds and verify identity or canonical structural equality with the selected turn entry. Reject mismatches before reading prohibitions, or accept only the turn and index and resolve the proposition internally.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/perception/src/proposals.ts` around lines 453-469, The new index option is trusted without checking its range or confirming that turn.propositions[index] matches the supplied proposition. The code uses that index to look up turn-wide prohibitions, while it constructs the command from the separate proposition argument. Require an integer index within bounds and verify identity or canonical structural equality with the selected turn entry. Reject mismatches before reading prohibitions, or accept only the turn and index and resolve the proposition internally. Reproduce R5 using `/tmp/circe-pr2-evidence/probes.test.ts` at PR #2 head 91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893. Keep valid requests working and run the full suite and typecheck after the change.", "followUp": false, "source": "453: // Prefer an explicit position; fall back to identity, then to a unique\n454: // structural match so rehydrated turns are not silently dropped.\n455: let index = options.index ?? turn.propositions.indexOf(proposition);\n456: if (index < 0) {\n457: const serialized = stableStringify(proposition);\n458: const matches = turn.propositions\n459: .map((candidate, candidateIndex) => ({ candidate, candidateIndex }))\n460: .filter(({ candidate }) => stableStringify(candidate) === serialized);\n461: if (matches.length === 1) index = matches[0]!.candidateIndex;\n462: }\n463: if (index < 0) return undefined;\n464: \n465: const issues = turn.propositions.map((candidate) => shapeIssue(candidate));\n466: const isAccepted = (candidate: SemanticProposition): boolean =>\n467: decisionsAccepted(candidate, options.acceptedDecisions);\n468: const analysis = analyzeTurn(turn, isAccepted, issues);\n469: if (analysis.prohibited.has(index) || analysis.unvalidatedProhibited.has(index)) {", "test": "R5", "commit": "91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893", "resolutionNote": "The original reproduction passes independently at 34685ba. Adjacent remaining paths and evidence regressions are recorded separately as F38-F39, without rewriting this finding." }, { "id": "f36", "status": "fixed", "fixedAt": "34685ba2cbb50d7d029437a040d1d3a6dfe0eac6", "reviewRound": 1, "title": "The per-proposition API drops malformed exclusions", "file": "packages/perception/src/proposals.ts", "lines": "465-468", "severity": "major", "pBadge": "P1", "type": "logic", "section": "actionable", "category": "correctness", "bodyParagraphs": [ "propositionCommand computes shape issues for every proposition but never rejects a malformed turn. analyzeTurn skips any proposition with a shape issue. In contrast, compileSemanticTurn rejects the whole turn before this step.", "The reproduction supplies an exclusion of A whose relations field is malformed, alongside a valid request for A. The whole-turn API correctly rejects it. The per-proposition API skips the malformed exclusion and emits the request for A, complete with current version evidence.", "The unvalidated-exclusion sets do not help here because the malformed proposition is skipped before any refs enter those sets. An API promising fail-closed exclusions must not turn a malformed constraint into permission." ], "fix": "Apply one whole-turn structural validation pass to both entry points. A malformed sibling must prevent command construction, or make the low-level helper internal to an already validated turn.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/perception/src/proposals.ts` around lines 465-468, propositionCommand computes shape issues for every proposition but never rejects a malformed turn. analyzeTurn skips any proposition with a shape issue. In contrast, compileSemanticTurn rejects the whole turn before this step. Apply one whole-turn structural validation pass to both entry points. A malformed sibling must prevent command construction, or make the low-level helper internal to an already validated turn. Reproduce R6 using `/tmp/circe-pr2-evidence/probes.test.ts` at PR #2 head 91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893. Keep valid requests working and run the full suite and typecheck after the change.", "followUp": false, "source": "465: const issues = turn.propositions.map((candidate) => shapeIssue(candidate));\n466: const isAccepted = (candidate: SemanticProposition): boolean =>\n467: decisionsAccepted(candidate, options.acceptedDecisions);\n468: const analysis = analyzeTurn(turn, isAccepted, issues);", "test": "R6", "commit": "91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893", "resolutionNote": "The original reproduction passes independently at 34685ba. Adjacent remaining paths and evidence regressions are recorded separately as F38-F39, without rewriting this finding." }, { "id": "f37", "status": "fixed", "fixedAt": "34685ba2cbb50d7d029437a040d1d3a6dfe0eac6", "reviewRound": 1, "title": "Frame binding checks content but omits identity", "file": "packages/perception/src/proposals.ts", "lines": "533-550", "severity": "minor", "pBadge": "P2", "type": "logic", "section": "actionable", "category": "data-integrity", "bodyParagraphs": [ "The whole-turn compiler checks frame.hash against turn.frameHash but does not compare frame.id with turn.frameId. propositionCommand checks both. Before the latest refactor the whole-turn compiler reached that identity check through the helper; it no longer does.", "The reproduction supplies the same content and hash under a different frame id. The turn still compiles. This is a provenance and contract inconsistency; the probe does not demonstrate different world content or an unsafe mutation, so it is lower severity than the execution defects.", "The accepted-turn contract and PR description record and promise both frame id and hash. Enforcing one of them in only one API makes the evidence-binding behavior depend on which compiler the caller chooses." ], "fix": "Require both frame id and hash to match in the shared validation path. If content equivalence is intentionally sufficient, make that explicit across the contract and both APIs instead of silently changing it in one.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/perception/src/proposals.ts` around lines 533-550, The whole-turn compiler checks frame.hash against turn.frameHash but does not compare frame.id with turn.frameId. propositionCommand checks both. Before the latest refactor the whole-turn compiler reached that identity check through the helper; it no longer does. Require both frame id and hash to match in the shared validation path. If content equivalence is intentionally sufficient, make that explicit across the contract and both APIs instead of silently changing it in one. Reproduce R7 using `/tmp/circe-pr2-evidence/probes.test.ts` at PR #2 head 91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893. Keep valid requests working and run the full suite and typecheck after the change.", "followUp": false, "source": "533: // Evidence must come from the frame the turn was decided against.\n534: if (options.frame && turn.frameHash === undefined) {\n535: return {\n536: commands: [],\n537: rejected: turn.propositions.map((_, index) => ({\n538: index,\n539: reason: \"turn has no frame identity to bind the supplied frame to\",\n540: })),\n541: };\n542: }\n543: if (options.frame && turn.frameHash !== undefined && options.frame.hash !== turn.frameHash) {\n544: return {\n545: commands: [],\n546: rejected: turn.propositions.map((_, index) => ({\n547: index,\n548: reason: \"frame does not match the frame the turn was decided against\",\n549: })),\n550: };", "test": "R7", "commit": "91c123b06b177c6b943fe2ac2f0ffd1e8c2f9893", "resolutionNote": "The original reproduction passes independently at 34685ba. Adjacent remaining paths and evidence regressions are recorded separately as F38-F39, without rewriting this finding." }, { "id": "f38", "status": "fixed", "fixedAt": "2cb82c8582ff2edf68fc8f47a5ec90d2aa1fada8", "reviewRound": 2, "title": "The per-proposition API still bypasses prerequisite rejection", "file": "packages/perception/src/proposals.ts", "lines": "521-527", "severity": "major", "pBadge": "P1", "type": "logic", "section": "actionable", "category": "correctness", "relatedFinding": "f33", "bodyParagraphs": [ "The new whole-turn compiler tracks rejectedIndexes and prevents a dependent proposition from compiling when its prerequisite is rejected. propositionCommand still checks only whether the prerequisite decision ids are accepted. Shared shape and act checks do not make these dependency rules shared.", "The reproduction uses an accepted request for B without a frame observation, followed by an accepted request for A with dependsOn:0. compileSemanticTurn correctly emits no commands. propositionCommand on the dependent request, with the correct explicit index, returns an executable command for A. This is the remaining entry-point path of F33; its original whole-turn reproduction now passes." ], "fix": "Share the complete dependency-validity analysis between both entry points, including evidence and prohibition rejection, or require the per-proposition helper to consume an already validated turn. Sharing only executableIssue is insufficient.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/perception/src/proposals.ts` around lines 521-527, The new whole-turn compiler tracks rejectedIndexes and prevents a dependent proposition from compiling when its prerequisite is rejected. propositionCommand still checks only whether the prerequisite decision ids are accepted. Shared shape and act checks do not make these dependency rules shared. Share the complete dependency-validity analysis between both entry points, including evidence and prohibition rejection, or require the per-proposition helper to consume an already validated turn. Sharing only executableIssue is insufficient. Reproduce N1 using `/tmp/circe-pr2-round2-evidence/round2.test.ts` at PR #2 head 34685ba2cbb50d7d029437a040d1d3a6dfe0eac6. Retain the 41 prior reviewer cases and the genuine argument-dependency control, and run the full suite and typecheck.", "followUp": false, "source": "521: const isAccepted = (candidate: SemanticProposition): boolean =>\n522: decisionsAccepted(candidate, options.acceptedDecisions);\n523: const analysis = analyzeTurn(turn, isAccepted, issues);\n524: if (analysis.prohibited.has(index) || analysis.unvalidatedProhibited.has(index)) {\n525: return undefined;\n526: }\n527: if (relationIssue(turn, proposition, index, isAccepted)) return undefined;", "test": "N1", "commit": "34685ba2cbb50d7d029437a040d1d3a6dfe0eac6", "resolutionNote": "The original reproduction passes independently at 2cb82c8. Newly introduced evidence-use regressions are recorded separately as F40-F41, without rewriting this finding." }, { "id": "f39", "status": "fixed", "fixedAt": "2cb82c8582ff2edf68fc8f47a5ec90d2aa1fada8", "reviewRound": 2, "title": "Slot evidence is incorrectly retained as a shared read dependency", "file": "packages/core/src/runtime.ts", "lines": "1088-1096", "severity": "major", "pBadge": "P1", "type": "logic", "section": "actionable", "category": "data-integrity", "relatedFinding": "f32", "bodyParagraphs": [ "applyProposal copies every command.expectedVersions entry into intent.readDependencies, including refs already owned by slot bindings. Revisions update the bindings but preserve this second copy through the intent spread. commit deduplicates with binding versions winning, while effect construction at lines 896-899 reverses that precedence and lets readDependencies win.", "The first reproduction proposes A at version 1, revises its binding to version 2 and confirms. Authority validation receives version 2; the executor receives version 1. A second reproduction retargets the intent from A to B. A is no longer a target, but its retained version still reaches validation and blocks B when A has changed.", "The same copy also breaks independent target effects. Every effect receives the entire readDependencies list, so an A/B request with maxConcurrentEffects:1 gives the B invocation a precondition on A. An executor that atomically checks preconditions updates A first, then refuses B because A is now newer. Only A is applied. All three evidence tests pass at 91c123b and fail at 34685ba; these are regressions introduced by the evidence repair.", "The shared argument-evidence fix itself works: a separate passing control verifies that a genuine argument ref C survives into both effects. The repair must retain that behavior while distinguishing slot-owned evidence from true shared dependencies." ], "fix": "Represent slot-owned expectations and shared read dependencies separately at the command boundary. Derive a canonical evidence set for each intent revision, validate and freeze the same versions, retire evidence owned only by replaced bindings, and attach only genuine shared dependencies to every target effect. Preserve argument dependencies; do not remove all shared evidence as a workaround.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/core/src/runtime.ts` around lines 1088-1096, applyProposal copies every command.expectedVersions entry into intent.readDependencies, including refs already owned by slot bindings. Revisions update the bindings but preserve this second copy through the intent spread. commit deduplicates with binding versions winning, while effect construction at lines 896-899 reverses that precedence and lets readDependencies win. Represent slot-owned expectations and shared read dependencies separately at the command boundary. Derive a canonical evidence set for each intent revision, validate and freeze the same versions, retire evidence owned only by replaced bindings, and attach only genuine shared dependencies to every target effect. Preserve argument dependencies; do not remove all shared evidence as a workaround. Reproduce N2, N3, N4 using `/tmp/circe-pr2-round2-evidence/round2.test.ts` at PR #2 head 34685ba2cbb50d7d029437a040d1d3a6dfe0eac6. Retain the 41 prior reviewer cases and the genuine argument-dependency control, and run the full suite and typecheck.", "followUp": false, "source": "1088: // The declared read set is command-wide: refs carried in arguments (or\n1089: // otherwise not bound to a slot) must still be validated and frozen.\n1090: const readDependencies = dedupePreconditions(command.expectedVersions ?? []);\n1091: const intent: Intent = {\n1092: id: `intent:${++state.counters.intent}`,\n1093: revision: 1,\n1094: capability: command.capability,\n1095: bindings,\n1096: ...(readDependencies.length > 0 ? { readDependencies } : {}),", "test": "N2, N3, N4", "commit": "34685ba2cbb50d7d029437a040d1d3a6dfe0eac6", "resolutionNote": "The original reproduction passes independently at 2cb82c8. Newly introduced evidence-use regressions are recorded separately as F40-F41, without rewriting this finding." }, { "id": "f40", "status": "fixed", "fixedAt": "a5d74b5790867568948c579ec07e2756fd6292fc", "reviewRound": 3, "title": "A ref used as both a target and an argument loses effect-level coverage", "file": "packages/core/src/runtime.ts", "lines": "1121-1125", "severity": "major", "pBadge": "P1", "type": "logic", "section": "actionable", "category": "data-integrity", "relatedFinding": "f39", "bodyParagraphs": [ "The new ownership filter removes every bound ref from shared readDependencies. A ref can simultaneously be a target and a shared argument dependency. When the runtime splits a command into target effects, each effect keeps the same arguments but only its own binding expectations.", "The reproduction compiles targets A and B with arguments.secondary = B. Both refs have frame evidence. After dispatch, the A invocation carries the B argument but has no precondition for B. Initial validation included B, but the executor cannot conditionally protect that read when it applies the A effect. This is a regression from 34685ba, where the same test passes.", "Canonical storage of one version per ref is reasonable. It does not mean the ref is used in only one role or by only one effect. The preceding tests covered shared argument C outside the target set and therefore missed this overlap." ], "fix": "Track evidence usage separately from its canonical version. A shared argument dependency must be attached to every effect that consumes it even when the same ref is also bound as a target. Preserve the passing independent-target case where there is no actual shared dependency.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/core/src/runtime.ts` around lines 1121-1125, The new ownership filter removes every bound ref from shared readDependencies. A ref can simultaneously be a target and a shared argument dependency. When the runtime splits a command into target effects, each effect keeps the same arguments but only its own binding expectations. Track evidence usage separately from its canonical version. A shared argument dependency must be attached to every effect that consumes it even when the same ref is also bound as a target. Preserve the passing independent-target case where there is no actual shared dependency. Reproduce O1 using `/tmp/circe-pr2-round3-evidence/ownership.test.ts` at PR #2 head 2cb82c8582ff2edf68fc8f47a5ec90d2aa1fada8. Retain all 46 prior reviewer cases, including independent targets and genuine shared dependencies, and run the full suite and typecheck.", "followUp": false, "source": "1121: const boundKeys = new Set(bindings.flatMap((binding) => binding.refs).map(entityRefKey));\n1122: const readDependencies = dedupePreconditions(\n1123: (command.expectedVersions ?? []).filter(\n1124: (precondition) => !boundKeys.has(entityRefKey(precondition.ref)),\n1125: ),", "test": "O1", "commit": "2cb82c8582ff2edf68fc8f47a5ec90d2aa1fada8", "resolutionNote": "All original F40-F41 probes now pass independently. Versions and read usage are separated, guards survive revision, and evidence transfers between roles. Original prose preserved." }, { "id": "f41", "status": "fixed", "fixedAt": "a5d74b5790867568948c579ec07e2756fd6292fc", "reviewRound": 3, "title": "Revisions silently discard still-required read dependencies", "file": "packages/core/src/runtime.ts", "lines": "1212-1231", "severity": "major", "pBadge": "P1", "type": "logic", "section": "actionable", "category": "data-integrity", "relatedFinding": "f39", "bodyParagraphs": [ "applyRevision now rebuilds readDependencies exclusively from refs found inside current arguments and constraints. It does not preserve an explicit command-wide read guard that has no corresponding argument. It also searches only the old shared list and latestSnapshots, ignoring evidence previously owned by a slot that the revision replaces. Missing versions are filtered out, after which commit checks coverage only for bound refs.", "One reproduction proposes A with an explicit read dependency on C, changes C to a newer version, then revises only a scalar note. C disappears from validation and the request invokes despite the stale declared guard. A second starts with B as both target and argument, then retargets to A while retaining the B argument. The old B version exists on the replaced binding but is not transferred. With no kernel snapshot, B disappears from the read set and the request invokes even though B is stale.", "Both cases pass on 34685ba and fail on 2cb82c8. The fix for obsolete slot evidence should retire evidence whose use ended, rather than discard explicit shared guards or evidence whose reference still appears in an executable argument." ], "fix": "Make revision update evidence from explicit dependency roles and the prior canonical evidence set, including replaced bindings. Preserve declared shared guards unless explicitly replaced or retired. Transfer versions for refs whose role changes, and reject missing required evidence instead of filtering it out. Validate coverage against the refs and guards the revised effect actually consumes.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/core/src/runtime.ts` around lines 1212-1231, applyRevision now rebuilds readDependencies exclusively from refs found inside current arguments and constraints. It does not preserve an explicit command-wide read guard that has no corresponding argument. It also searches only the old shared list and latestSnapshots, ignoring evidence previously owned by a slot that the revision replaces. Missing versions are filtered out, after which commit checks coverage only for bound refs. Make revision update evidence from explicit dependency roles and the prior canonical evidence set, including replaced bindings. Preserve declared shared guards unless explicitly replaced or retired. Transfer versions for refs whose role changes, and reject missing required evidence instead of filtering it out. Validate coverage against the refs and guards the revised effect actually consumes. Reproduce O2, O3 using `/tmp/circe-pr2-round3-evidence/ownership.test.ts` at PR #2 head 2cb82c8582ff2edf68fc8f47a5ec90d2aa1fada8. Retain all 46 prior reviewer cases, including independent targets and genuine shared dependencies, and run the full suite and typecheck.", "followUp": false, "source": "1212: const existingReadDependencies = new Map(\n1213: (intent.readDependencies ?? []).map((precondition) => [\n1214: entityRefKey(precondition.ref),\n1215: precondition,\n1216: ]),\n1217: );\n1218: const sharedRefs = refsInValues([argumentsNow, intent.constraints]).filter(\n1219: (ref) => !boundKeys.has(entityRefKey(ref)),\n1220: );\n1221: const observed = latestSnapshots(sharedRefs);\n1222: const readDependencies = dedupePreconditions(\n1223: sharedRefs.map((ref) => {\n1224: const existing = existingReadDependencies.get(entityRefKey(ref));\n1225: if (existing) return existing;\n1226: const snapshot = observed.get(entityRefKey(ref));\n1227: if (snapshot && snapshot.version.authority === ref.authority) {\n1228: return { ref: { ...ref }, token: { ...snapshot.version } };\n1229: }\n1230: return undefined;\n1231: }).filter((precondition): precondition is FrozenPrecondition => Boolean(precondition)),", "test": "O2, O3", "commit": "2cb82c8582ff2edf68fc8f47a5ec90d2aa1fada8", "resolutionNote": "All original F40-F41 probes now pass independently. Versions and read usage are separated, guards survive revision, and evidence transfers between roles. Original prose preserved." }, { "id": "f42", "status": "fixed", "fixedAt": "ae4ae3de85905bc2f7c2752679168991cf51561c", "reviewRound": 4, "title": "An unrelated old revision permanently blocks valid new continuations", "file": "packages/core/src/runtime.ts", "lines": "1408-1417", "severity": "minor", "pBadge": "P2", "type": "logic", "section": "actionable", "category": "correctness", "bodyParagraphs": [ "applyProposal now compares expectedIntentRevision against every intent ever stored in the interaction. Revision numbers belong to individual intents and restart at 1. A larger number on an unrelated intent is not evidence that the presented intent has been superseded.", "The reproduction proposes A, revises it to revision 2, and cancels it. It then proposes B at revision 1 and submits a continuation of B with expectedIntentRevision 1. The runtime returns STALE_INTENT even though B is the presented live intent and its revision matches exactly. The old cancelled A remains in state.intents, so the same condition poisons subsequent fresh revision-1 continuations for the rest of this interaction.", "This is a functional regression, not a reproduced unauthorized dispatch. It passes at 2cb82c8 and fails at a5d74b5 and stack tip 8cea673. A control without the unrelated old revision still succeeds. The current tests exercise stale continuations but omit this valid continuation after an unrelated transaction ends." ], "fix": "Bind continuation evidence to an intent or transaction identity plus its revision. Compare that named anchor, including its lifecycle, rather than the maximum revision anywhere in the interaction. Preserve rejection when the named intent was revised, replaced or settled; simply deleting the global check would reopen the stale-anchor problem.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/core/src/runtime.ts` around lines 1408-1417, applyProposal now compares expectedIntentRevision against every intent ever stored in the interaction. Revision numbers belong to individual intents and restart at 1. A larger number on an unrelated intent is not evidence that the presented intent has been superseded. Bind continuation evidence to an intent or transaction identity plus its revision. Compare that named anchor, including its lifecycle, rather than the maximum revision anywhere in the interaction. Preserve rejection when the named intent was revised, replaced or settled; simply deleting the global check would reopen the stale-anchor problem. Reproduce using `/tmp/circe-pr2-round4-evidence/controls.test.ts`, then retain all 49 earlier review cases and both new positive controls. Run the baseline suite and typecheck. This review covers merged PR #2 at a5d74b5790867568948c579ec07e2756fd6292fc; current tip also reproduces the failure.", "followUp": false, "source": "1408: const declared = command.expectedIntentRevision;\n1409: const superseded = [...state.intents.values()].some(\n1410: (intent) => intent.revision > declared,\n1411: );\n1412: const anchor = presentedTransaction();\n1413: const anchorIntent = anchor ? state.intents.get(anchor.intentId) : undefined;\n1414: const stale =\n1415: superseded ||\n1416: (anchorIntent !== undefined && anchorIntent.revision !== declared) ||\n1417: (anchorIntent === undefined && state.hasPresentedGate);", "test": "P1", "commit": "a5d74b5790867568948c579ec07e2756fd6292fc", "resolutionNote": "All original round-4 probes pass independently at ae4ae3d. Named-current continuation after an unrelated completed revision also passes. New identity integration gaps are recorded as F44-F45; original prose preserved." }, { "id": "f43", "status": "fixed", "fixedAt": "ae4ae3de85905bc2f7c2752679168991cf51561c", "reviewRound": 4, "title": "The compiler invents slot mechanics that disagree with the kernel", "file": "packages/perception/src/proposals.ts", "lines": "532-550", "severity": "minor", "pBadge": "P2", "type": "logic", "section": "actionable", "category": "correctness", "bodyParagraphs": [ "The compiler chooses the first semantic overlay slot as the default target and assumes every other slot is scalar unless the overlay repeats its arity. The kernel instead prefers the mechanical targetSlot, then a slot named targets, then a required slot. Mechanical arities already exist on CapabilityDescriptor. The semantic overlay fields are optional, and validateCapabilityOverlays accepts overlays without them.", "With a valid overlay ordered as optional note followed by targets, a proposition whose summary refs contain A now compiles to note=A instead of targets=A. A second case names targets=A and watchers=[B,C], with watchers mechanically declared many and its optional semantic arity omitted. The kernel accepts this command, but the compiler rejects it as a scalar-slot violation. Both overlays pass the public overlay validator.", "Both cases pass at 2cb82c8 and fail at the merged PR head and current stack tip. Supplying complete semantic arity makes the second control pass. These are valid-command regressions caused by duplicating the execution schema in the perception layer; neither probe demonstrates an unauthorized external effect." ], "fix": "Use one authoritative mechanical slot contract. Materialize normalized targetSlot, requiredness and arity from CapabilityDescriptor into the frame, or share the normalization rule without making core depend on perception. Do not interpret missing optional semantic metadata as a contradictory execution constraint. Verify default-target behavior under slot reordering and valid secondary many-valued slots.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/perception/src/proposals.ts` around lines 532-550, The compiler chooses the first semantic overlay slot as the default target and assumes every other slot is scalar unless the overlay repeats its arity. The kernel instead prefers the mechanical targetSlot, then a slot named targets, then a required slot. Mechanical arities already exist on CapabilityDescriptor. The semantic overlay fields are optional, and validateCapabilityOverlays accepts overlays without them. Use one authoritative mechanical slot contract. Materialize normalized targetSlot, requiredness and arity from CapabilityDescriptor into the frame, or share the normalization rule without making core depend on perception. Do not interpret missing optional semantic metadata as a contradictory execution constraint. Verify default-target behavior under slot reordering and valid secondary many-valued slots. Reproduce using `/tmp/circe-pr2-round4-evidence/controls.test.ts`, then retain all 49 earlier review cases and both new positive controls. Run the baseline suite and typecheck. This review covers merged PR #2 at a5d74b5790867568948c579ec07e2756fd6292fc; current tip also reproduces the failure.", "followUp": false, "source": "532: const fallbackSlot =\n533: offeredCapability?.targetSlot ??\n534: offeredCapability?.slots[0]?.name ??\n535: \"targets\";\n536: const declaredSlots = new Map(\n537: (offeredCapability?.slots ?? []).map((slot) => [slot.name, slot]),\n538: );\n539: if (offeredCapability && offeredCapability.slots.length > 0) {\n540: const unknownSlot = (proposition.slots ?? []).find(\n541: (slot) => !declaredSlots.has(slot.slot),\n542: );\n543: if (unknownSlot) {\n544: return fail(\n545: `slot ${unknownSlot.slot} is not offered by capability ${proposition.capability}`,\n546: );\n547: }\n548: }\n549: const effectiveArity = (slotName: string): \"one\" | \"many\" =>\n550: declaredSlots.get(slotName)?.arity ?? (slotName === fallbackSlot ? \"many\" : \"one\");", "test": "P2, P3", "commit": "a5d74b5790867568948c579ec07e2756fd6292fc", "resolutionNote": "All original round-4 probes pass independently at ae4ae3d. Named-current continuation after an unrelated completed revision also passes. New identity integration gaps are recorded as F44-F45; original prose preserved." }, { "id": "f44", "status": "fixed", "fixedAt": "672b86d991b1f23d6a7dc5fc4be38542ce59a0e7", "reviewRound": 5, "title": "Semantic ingestion drops the named continuation guard", "file": "packages/core/src/runtime.ts", "lines": "1839-1843", "severity": "major", "pBadge": "P1", "type": "logic", "section": "actionable", "category": "correctness", "bodyParagraphs": [ "The new expectedIntentId check applies only if the ID reaches applyProposal. proposalFromFrame forwards expectedIntentRevision but does not copy expectedIntentId, and the internal SemanticFrame shape does not declare it. A semantic classifier can return a named continuation that becomes unnamed at the consequence boundary.", "The reproduction cancels intent A at revision 1 and presents intent B at revision 1. A typed proposal carrying expectedIntentId=A is rejected as STALE_INTENT. Ingesting the identical resolved fields drops A's identity, validates against B's matching revision, and creates a fresh A proposal. Confirming that proposal invokes A. The test demonstrates the stale named continuation being admitted through ingestion; the invocation still requires the normal confirmation in this fixture.", "This defeats the newly introduced named-anchor lifecycle check on an existing entry point. Preserve the identity through the bridge rather than relying on the unrelated presented gate to supply it." ], "fix": "Declare and decode expectedIntentId in SemanticFrame and forward it into the detached ProposeCommand. Fail closed for a present but malformed identity rather than silently converting it to an unnamed continuation. Retain equivalent named stale/current tests for command() and ingest().", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/core/src/runtime.ts` around lines 1839-1843, The new expectedIntentId check applies only if the ID reaches applyProposal. proposalFromFrame forwards expectedIntentRevision but does not copy expectedIntentId, and the internal SemanticFrame shape does not declare it. A semantic classifier can return a named continuation that becomes unnamed at the consequence boundary. Declare and decode expectedIntentId in SemanticFrame and forward it into the detached ProposeCommand. Fail closed for a present but malformed identity rather than silently converting it to an unnamed continuation. Retain equivalent named stale/current tests for command() and ingest(). Reproduce Q1 using `/tmp/circe-pr2-round5-evidence/identity.test.ts` at PR #7 head ae4ae3de85905bc2f7c2752679168991cf51561c. Retain all 54 previous probes and the named-current-intent control. Run the full suite and typecheck; keep this a narrow identity propagation fix.", "followUp": false, "source": "1839: ...(expectedVersions !== undefined ? { expectedVersions } : {}),\n1840: ...(typeof frame.expectedIntentRevision === \"number\"\n1841: ? { expectedIntentRevision: frame.expectedIntentRevision }\n1842: : {}),\n1843: provenance: classification.provenance,", "test": "Q1", "commit": "ae4ae3de85905bc2f7c2752679168991cf51561c", "resolutionNote": "Q1 now rejects the cancelled named continuation through semantic ingestion; valid named ingestion and malformed-ID controls also pass." }, { "id": "f45", "status": "fixed", "fixedAt": "672b86d991b1f23d6a7dc5fc4be38542ce59a0e7", "reviewRound": 5, "title": "Command deduplication ignores the new continuation identity", "file": "packages/perception/src/proposals.ts", "lines": "465-474", "severity": "minor", "pBadge": "P2", "type": "logic", "section": "actionable", "category": "correctness", "bodyParagraphs": [ "buildCommand now preserves targetIntentId as expectedIntentId, but dedupeKeyOf still compares only the expected revision alongside the executable payload. Two different intent anchors can have the same local revision counter. They impose different lifecycle and revision checks and must not collapse into one command.", "The reproduction compiles two otherwise identical propositions at revision 1. The first names a cancelled intent; the second names the current live intent. Compilation emits only the first ID and labels the second proposition a duplicate. The retained stale command is rejected by the kernel, so the valid continuation is lost. Reversing the order changes which anchor survives.", "Include expectedIntentId in the canonical key, just as expectedIntentRevision is already included. Keep deduplication for propositions whose payload and full continuation guard actually match." ], "fix": "Add expectedIntentId to dedupeKeyOf. Test otherwise identical propositions with different named anchors, including a stale-first/current-second case, and preserve a same-anchor duplicate control.", "prompt": "Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate.\n\nIn `packages/perception/src/proposals.ts` around lines 465-474, buildCommand now preserves targetIntentId as expectedIntentId, but dedupeKeyOf still compares only the expected revision alongside the executable payload. Two different intent anchors can have the same local revision counter. They impose different lifecycle and revision checks and must not collapse into one command. Add expectedIntentId to dedupeKeyOf. Test otherwise identical propositions with different named anchors, including a stale-first/current-second case, and preserve a same-anchor duplicate control. Reproduce Q2 using `/tmp/circe-pr2-round5-evidence/identity.test.ts` at PR #7 head ae4ae3de85905bc2f7c2752679168991cf51561c. Retain all 54 previous probes and the named-current-intent control. Run the full suite and typecheck; keep this a narrow identity propagation fix.", "followUp": false, "source": "465: return stableStringify({\n466: capability: command.capability,\n467: slots: command.slots\n468: .map((slot) => ({ slot: slot.slot, refs: canonicalRefKeys(slot.refs) }))\n469: .sort((a, b) => (a.slot < b.slot ? -1 : a.slot > b.slot ? 1 : 0)),\n470: argumentRefs: canonicalRefKeys(argumentRefsOfCommand(command)),\n471: arguments: command.arguments ?? null,\n472: constraints: command.constraints ?? null,\n473: expectedIntentRevision: command.expectedIntentRevision ?? null,\n474: });", "test": "Q2", "commit": "ae4ae3de85905bc2f7c2752679168991cf51561c", "resolutionNote": "Q2 now preserves distinct named anchors, rejects the stale one and admits the current one. Retained repository test also confirms same-anchor duplicates still collapse." } ], "history": [ { "round": 1, "at": "2026-09-22T12:41:53.536939+00:00", "actionableOpen": 7, "verdict": "CHANGES_REQUESTED", "note": "Reviewed PR #2 after PR #1 merge at exact head 91c123b. Findings continue the earlier review numbering as F31-F37. No Circe source edits." }, { "round": 2, "at": "2026-09-22T12:59:04.708674+00:00", "actionableOpen": 2, "verdict": "CHANGES_REQUESTED", "note": "Original F31-F37 cases verified fixed. Added F38 for the remaining dependency entry point and F39 for duplicated slot evidence across revisions and effects. No source edits." }, { "round": 3, "at": "2026-09-22T17:13:11.179624+00:00", "actionableOpen": 2, "verdict": "CHANGES_REQUESTED", "note": "Original F38-F39 cases verified fixed. Appended F40-F41 for evidence shared between roles and still-required read dependencies dropped by revision. No source edits." }, { "round": 4, "at": "2026-09-22T20:51:05.040219+00:00", "actionableOpen": 2, "verdict": "COMMENT", "note": "F40-F41 independently verified fixed. Appended F42-F43, two P2 functional regressions. Assessed added complexity and recommended narrow contract fixes, with no rewrite or revert. Circe source untouched." }, { "round": 5, "at": "2026-09-22T21:05:35.964356+00:00", "actionableOpen": 2, "verdict": "CHANGES_REQUESTED", "note": "Marked original F42-F43 cases fixed. Added F44 for named guard lost in ingestion and F45 for named guards conflated by deduplication. No Circe source edits." }, { "round": 6, "at": "2026-09-22T21:14:18.136742+00:00", "actionableOpen": 0, "verdict": "APPROVED", "note": "F44-F45 independently verified fixed. No new findings. Review recommendation only; no GitHub approval or merge submitted." } ], "syncMeta": { "instructions": "Preserve baseline and finding prose on later reviews. Review the exact PR #2 head and append only new findings. PR #1 approval remains separate." } }