enhance(verify-phase): deterministic auto-locate of the prohibition check descriptor (#1278) (#1301)

* test(1278): RED-first descriptor parity + fail-closed guards + CHK-07 byte-stability (wave 1)

- CHK-03 (RED): extend PROB-14 parity in prohibition-probe.schema.test.cjs to carry the flat
  check_kind/check_target/check_rule scalars through project->write->parseMustHavesBlock; the
  non-droppable check_kind-presence assertion is the load-bearing RED trigger (fails because
  projectProhibitions strips check_* on the current build).
- CHK-07 (GREEN forward-guard): probe-core.test.cjs pins descriptor-less byte-stability +
  dispositionForProhibition fail-closed policy, with a t.todo marker forward-locking plan 01-02.
- CHK-06 (RED): prohibition-enforcement.test.cjs asserts descriptorFromProjection export +
  fail-closed on absent/partial/unknown descriptors via the projection adapter (RED until 01-03).
- No src/*.cts or .cjs edits; no new test files; lint-test-file-count clean.

* feat(1278): add optional flat-scalar check descriptor fields to Prohibition interface (wave 2)

- check_kind?/check_target?/check_rule? mirror CheckDescriptor.kind/target/rule (minus caller-attested failFirst, #1279)
- optional so existing Prohibition consumers compile unchanged

* feat(1278): project check descriptor as flat scalars in projectProhibitions (wave 2)

- emit check_kind/check_target (+ check_rule only for lint-rule with a rule) when descriptor well-formed
- under-specified/descriptor-less items project byte-identically (CHK-07); flat scalars ride existing parseMustHavesBlock continuation-KV path (no parser rewrite)
- add CHK-02 probe-core unit cases pinning the projection
- turns CHK-03 parity test GREEN; dispositionForProhibition untouched

* feat(1278): descriptorFromProjection read-back adapter feeds fail-closed locate (wave 3)

- Add descriptorFromProjection(projected) -> CheckDescriptor | null to
  src/prohibition-enforcement.cts: renames the projected flat scalars
  check_kind/check_target/check_rule -> {kind,target,rule?}, or null when
  the descriptor is absent/non-object (no check_kind key).
- failFirst is NEVER sourced from the projection (stays caller-attested; #1279).
- rule is set only when check_rule is a non-empty string; the adapter does NOT
  re-validate kind/target/rule — an under-specified descriptor reconstructs to
  one the EXISTING runProhibitionEnforcement LOCATE guard rejects (located:false,
  never green). The merged #1259 guard stays the single source of fail-closed truth.
- Turns the RED CHK-06 fail-closed tests (plan 01-01) GREEN end-to-end; CHK-03 /
  CHK-07 stay green. CheckDescriptor type, locate guard, dispositionForProhibition,
  and parseMustHavesBlock are unchanged (additive +36/-0).

* feat(1278): verify-phase locates prohibition check from projected descriptor (wave 3)

- request.check kind/target/rule sourced from projected check_kind/check_target/check_rule via descriptorFromProjection, not verifier invention (CHK-05)
- replaces the #1278 author-supplied / tracked-follow-up note with the delivered deterministic-locate behavior
- preserves fail-closed routing: absent/partial descriptor -> never green, hard-gate in both modes
- failFirst stays a verify-time caller attestation; #1279 bounds the remaining fail-first proof

* feat(1278): spec-phase captures wired-check descriptor on test-tier resolution (wave 3)

- Step 5.6 'Keep it' / verification: test path captures check_kind/check_target/check_rule, projected onto must_haves.prohibitions for verify-phase deterministic locate (CHK-04)
- SOFT capture: a test-tier prohibition without a descriptor is still allowed (no hard authoring block); stays fail-closed/flagged downstream
- --auto captures only an unambiguous descriptor, never fabricates a check path
- failFirst NOT captured at spec-phase (verify-time attestation; #1279)
- PROB-06 soft-gate + text-mode (PROB-09) behavior unchanged

* chore(1278): re-baseline workflow size for grown verify-phase + spec-phase prose (wave 3)

- spec-phase.md 28438 -> 30343 (+1905), verify-phase.md 35362 -> 36498 (+1136)
- regenerated via npm run size:baseline (no hand-picked numbers); growth is the #1278 deterministic-locate + descriptor-capture prose
- workflow-size-budget guard green (122/122)

* docs(1278): ratify optional check descriptor in dated ADR-550 addendum + type:Changed changeset

- Append dated 2026-06-15 ADR-550 addendum ratifying the D3 prohibition-item
  shape extension (optional flat-scalar check_kind/check_target/check_rule)
- Document flat-scalar rationale, deterministic projection/read-back,
  fail-closed on partial/invalid/absent, #1279/policy out-of-scope
- Add .changeset/1278-prohibition-check-descriptor.md (type: Changed)

* docs(1278): document optional check descriptor in prohibition-probe reference + FEATURES

- Add 'Optional wired-check descriptor (deterministic locate, #1278)' section
  to the prohibition-probe reference (flat-scalar keys, projection/read-back,
  fail-closed + backward-compat, failFirst stays attested)
- Add deterministic prohibition-check descriptor source entry to FEATURES.md
- No CONTEXT.md glossary change: descriptor reuses existing wired-check /
  verification:test vocabulary, no new glossary term introduced

* fix(1278): pass packaging gates — changeset pr field + retired slash-form fix

- Add required pr: 1278 to changeset (lint:changeset MISSING_PR hard requirement;
  plan's 'omit if unknown' was inaccurate — issue number per #1259 convention,
  updated to real PR number when opened) [Rule 3 - blocking]
- Fix retired /gsd-spec-phase -> /gsd:spec-phase at verify-phase.md:83 (wave-3
  prose; caught by slash-namespace invariant #3443/bug-2543, blocked CHK-09
  full-suite-green) [Rule 1 - bug]
- size:baseline + INVENTORY manifest verified in-sync post-build (no diff)

* docs(1278): add check descriptor + descriptorFromProjection to CONTEXT.md prohibition glossary

* fix(1278): harden descriptorFromProjection round-trip (numeric-coercion + stray-rule) per review

- MD-01/LW-01: narrow projected scalars to primitives + String()-coerce, so a
  numeric-looking check_target (parseMustHavesBlock coerces ^\d+$ to number)
  reconstructs as a string and locates instead of silently un-locating; no
  as-string type-lie, satisfies no-base-to-string.
- LW-02: attach rule only for the lint-rule kind (drop a stray node-test rule).
- LW-03: document the optional check_* keys in the reference Output schema.
RED->GREEN tests added in prohibition-enforcement.test.cjs.

* chore(1278): set changeset pr to 1301

* test(1278): add fast-check property for the check-descriptor round-trip + fail-closed (trek-e review)

RULESET.TESTS.property-based-testing: the projectProhibitions -> render ->
parseMustHavesBlock -> descriptorFromProjection chain is a bijective/transformation
contract. Adds 2 fc properties to tests/probe-core.property.test.cjs (no new file;
ratchet stays at 2 for probe-core):
- well-formed descriptors survive the round-trip across the full string domain
  incl. the numeric-coercion case (target/rule reconstruct as strings);
- under-specified/invalid descriptors (absent / target-less / rule-less /
  unknown-kind) are always fail-closed (never green, flagged, unlocated).
Stability is asserted at the descriptorFromProjection layer (the raw parse step is
intentionally lossy for numeric scalars; the shared parser is unchanged).

---------

Co-authored-by: Tom Boucher <trekkie@nomorestars.com>
This commit is contained in:
Rezolv
2026-06-16 00:01:03 -04:00
committed by GitHub
parent e3b829e765
commit e50ead7ad2
14 changed files with 584 additions and 10 deletions

View File

@@ -356,6 +356,21 @@ For each Requirement gathered so far, run the two-stage recall→precision pass:
Criteria AND mark the prohibition `resolved` with a verification tier: `test` (a
mechanical negative test/lint/assertion exists) or `judgment` (real but not mechanically
checkable — routes to judgment review).
- **Capture the wired-check descriptor on `test`-tier (#1278, SOFT).** When a prohibition is
resolved `verification: test`, ALSO capture the descriptor of the wired check so
`verify-phase` can LOCATE it deterministically (no verifier invention at verify time).
Capture the flat scalars — persisted into SPEC and projected onto the
`must_haves.prohibitions` item by `projectProhibitions`:
- `check_kind` — `node-test` | `lint-rule`.
- `check_target` — the negative-test file path (for `node-test`), or the path to lint
(for `lint-rule`).
- `check_rule` — the eslint rule id (e.g. `local/no-source-grep`); `lint-rule` only.
This is a **SOFT capture (CHK-04): a `test`-tier prohibition WITHOUT a descriptor is still
allowed** — if the author cannot yet name the wired check, leave the descriptor empty and
proceed. It is NOT a hard authoring block; the item simply stays fail-closed/flagged
downstream (an absent/partial descriptor → `descriptorFromProjection` null/under-specified
→ producer fail-closed locate, never green). Do NOT capture `failFirst` here — it is a
verify-time caller attestation, not a spec-authored field (#1279).
- **Dismiss (reason)** → mark `dismissed` with a REQUIRED non-empty reason (PROB-05). The
reason string is the audit trail; silence is not a valid dismissal.
- **Defer** → leave `unresolved`.
@@ -374,7 +389,11 @@ For each Requirement gathered so far, run the two-stage recall→precision pass:
**`--auto` mode:** auto-`resolved` where a defensible negative acceptance criterion can be
written (test or judgment tier); otherwise leave `unresolved`. **`--auto` NEVER auto-dismisses
a prohibition** — a wrong dismissal is the exact silent failure this probe eliminates (PROB-06,
the load-bearing safety property). Log: `[auto] prohibitions: R resolved, U unresolved`.
the load-bearing safety property). On a `test`-tier auto-resolution, capture the `check_kind` /
`check_target` / `check_rule` descriptor **only when a wired check is unambiguous**; otherwise
leave it empty — `--auto` NEVER fabricates a check path (a wrong locate is re-validated and
fails closed at the producer, but a fabricated path is still noise to avoid). Log:
`[auto] prohibitions: R resolved, U unresolved`.
**Text mode (PROB-09):** per Step 5's text-mode rule, replace the AskUserQuestion menus above
with plain-text numbered lists — there is NO hard AskUserQuestion dependency, so the probe
@@ -382,7 +401,11 @@ runs identically for non-Claude / text-mode hosts.
Populate the `## Prohibitions` section of SPEC.md from the resolved prohibitions (each
`resolved`/`test` row is a checkable negative acceptance criterion; `resolved`/`judgment`
rows route to judgment review; `⚠ UNRESOLVED` rows are flagged as assumptions).
rows route to judgment review; `⚠ UNRESOLVED` rows are flagged as assumptions). A
`resolved`/`test` row ALSO carries its captured `check_kind` / `check_target` / `check_rule`
descriptor when present (so the projection feeds `verify-phase`'s deterministic locate, #1278);
a `test` row with no captured descriptor is still valid — it stays fail-closed/flagged
downstream rather than blocking authoring.
## Step 6: Generate SPEC.md

View File

@@ -70,17 +70,17 @@ Aggregate all must_haves across plans for phase-level verification.
**Prohibitions (`must_haves.prohibitions`, ADR-550 D3 — the must-NOT sibling block):** When a plan carries `must_haves.prohibitions`, extract each `{ statement, status, verification }` item and route it by `verification` tier in verdict assembly (ADR-550 D4, "B-with-guard", 2026-06-12 maintainer decision). These are NEGATIVE checks (the must-NOT must NOT have happened), distinct from positive `truths`:
- **judgment-tier → mode-dependent soft-gate.** Interactive verify defers each item to the end-of-phase human checkpoint (`human_verify_mode: end-of-phase`). Autonomous verify records a NON-AUTHORITATIVE LLM-judge verdict + a prominent `unverified-prohibition — human review recommended` flag (autonomous completion reads "complete with N flagged prohibitions"). NEVER a silent pass; NEVER a hard halt of an AFK run.
- **test-tier → ENFORCED via `check prohibition-enforcement` (green on pass, hard-gate on miss/fail).** Accept the `verification: test` value (the SPEC↔must_haves.prohibitions projection contract holds — no forced schema change later). For each test-tier item, the verifier invokes the deterministic producer:
- **test-tier → ENFORCED via `check prohibition-enforcement` (green on pass, hard-gate on miss/fail).** Accept the `verification: test` value (the SPEC↔must_haves.prohibitions projection contract holds — no forced schema change later). For each test-tier item, the verifier builds `request.check` **DETERMINISTICALLY from the projected descriptor** — it does NOT invent `{ kind, target, rule }`. Read the flat scalar keys `check_kind` / `check_target` / `check_rule` off the `must_haves.prohibitions` item and reconstruct the `CheckDescriptor` via the `descriptorFromProjection` adapter in `prohibition-enforcement` (`descriptorFromProjection(projectedItem)` → `{ kind: check_kind, target: check_target, rule?: check_rule }`). Then attest `failFirst: true` in the request — `failFirst` is the ONE field NOT sourced from the projection; it stays a verify-time caller attestation (#1279 machine-proves it against a violation fixture). Invoke the producer (CLI surface unchanged):
```bash
gsd_run check prohibition-enforcement <request.json>
```
where `<request.json>` carries `{ prohibition, check, mode }` — `check` being the wired mechanical-check descriptor `{ kind: 'node-test' | 'lint-rule', target, rule?, failFirst: true }`. For `node-test`, `target` is the negative-test file path; for `lint-rule`, `target` is the PATH to lint and `rule` is the eslint rule id (e.g. `local/no-source-grep`) — both required (a lint-rule without `rule` is not a valid wired check). The producer LOCATES the wired check, requires the caller-attested `failFirst` marker, RUNS it for a genuine non-vacuous pass, builds `enforcementEvidence`, and emits the `dispositionForProhibition()` verdict (#1259, ADR-550 D5d). (`failFirst` is caller-attested, not yet machine-proven against a violation fixture — a tracked follow-up.) Route the result by its typed fields:
where `<request.json>` carries `{ prohibition, check, mode }` — `check` being the wired mechanical-check descriptor `{ kind: 'node-test' | 'lint-rule', target, rule?, failFirst: true }`, with `kind`/`target`/`rule` now sourced from the projected `check_*` scalars (not author/verifier invention) and `failFirst` caller-attested. For `node-test`, `target` (from `check_target`) is the negative-test file path; for `lint-rule`, `target` is the PATH to lint and `rule` (from `check_rule`) is the eslint rule id (e.g. `local/no-source-grep`) — both required (a lint-rule without `rule` is not a valid wired check). The producer LOCATES the wired check, requires the caller-attested `failFirst` marker, RUNS it for a genuine non-vacuous pass, builds `enforcementEvidence`, and emits the `dispositionForProhibition()` verdict (#1259, ADR-550 D5d). Route the result by its typed fields:
- **`status: 'green'`, `flagged: false`** (a genuinely-passing wired negative test / lint rule, `located: true`, non-empty `evidence`) → the item is satisfiable → it can reach **passed**.
- **missing, non-attested, or genuinely-non-passing check** (`located: false` OR `status: 'unverified'`, `flagged: true`) → **hard-gate**: disposes flagged-unverified, NEVER green, routing to `gaps_found` in BOTH interactive and autonomous modes (a failing mechanical check blocks even AFK; ADR-550 D4 / D3). The deterministic fail-closed default backing every miss/fail is `dispositionForProhibition()` in probe-core (`status: 'unverified'`, `flagged: true` on empty `enforcementEvidence`).
> **Descriptor authoring — current scope (#1259).** The `check` descriptor is **supplied by the phase author / verifier**; there is no projection field yet that deterministically derives `{ kind, target, rule, failFirst }` from a prohibition in `must_haves.prohibitions` (which carries only `{ statement, status, verification }`). So #1259 lands the **deterministic run+verdict half** (locate→run→evidence→disposition, all CI-testable) while the **locate→descriptor half is author-provided** for now. Deterministic auto-locate — so a wired passing test closes the gap with zero manual descriptor authoring — is a **tracked follow-up: #1278**. Until then, the green path requires the author to wire the descriptor explicitly.
> **Descriptor source — deterministic locate (#1278, DELIVERED).** The `check` descriptor's `{ kind, target, rule }` is now sourced **deterministically from the projected `check_kind` / `check_target` / `check_rule` scalars** on the `must_haves.prohibitions` item (authored at `/gsd:spec-phase`, projected by `projectProhibitions`, read back via the `descriptorFromProjection` adapter). So a wired passing test closes the gap with **zero manual descriptor authoring** — the verifier no longer invents the locate (removing the spoofable invent-at-verify-time surface; ADR-857 §147 exogenous grading). **Fail-closed is preserved:** an item with NO projected descriptor — or a PARTIAL one (e.g. a `lint-rule` missing `check_rule`) — makes `descriptorFromProjection` return `null` / an under-specified descriptor, which falls through to the producer's existing fail-closed LOCATE (`located: false`) → flagged-unverified, NEVER green, in BOTH modes. The only field still attested at verify time (not projected) is `failFirst`; machine-proving it against a violation fixture is the remaining **tracked follow-up: #1279**.
**Option B: Use Success Criteria from ROADMAP.md**