Epic: pipeline-integrity hardening tranche (handoff §3) — type gate, coverage ratchet, dep locking, adversarial procedure, break-it phase #147

Closed
opened 2026-09-05 19:22:56 +01:00 by conductor · 4 comments
Owner

Plain summary

Bounded five-item hardening tranche from RUNE_VIBE_CODE_INTEGRITY_HANDOFF §3 — engineering safeguards so LLM-authored changes that share one wrong assumption (implementation + tests + docs) are harder to merge. Explicitly NOT new cognitive architecture; NOT a permanent project — a finite tranche before P2 developmental-memory work (§9 Phase B).

Child issues

  1. Type-check dependency in the protected merge gate (§3.1)
  2. Coverage ratchet on changed/critical paths (§3.2)
  3. Reproducible dependency locking for CI (§3.3)
  4. Black-box adversarial test procedure for P0/P1 changes (§3.4)
  5. "Break-it" phase for constitutional paths (§3.5)

Anti-complexity rules (handoff §10, binding)

  • A concern is not automatically a subsystem. Prefer one invariant / one schema field / one trace relation / one test / one CI gate over a new service.
  • No future-scale machinery before current-scale evidence requires it.
  • Every mechanism must answer: trigger / state read / state changed / verification / failure mode / traceability / reversibility / benefit measurement / dormant cost / why an existing authority cannot own it.
## Plain summary Bounded five-item hardening tranche from RUNE_VIBE_CODE_INTEGRITY_HANDOFF §3 — engineering safeguards so LLM-authored changes that share one wrong assumption (implementation + tests + docs) are harder to merge. Explicitly NOT new cognitive architecture; NOT a permanent project — a finite tranche before P2 developmental-memory work (§9 Phase B). ## Child issues 1. Type-check dependency in the protected merge gate (§3.1) 2. Coverage ratchet on changed/critical paths (§3.2) 3. Reproducible dependency locking for CI (§3.3) 4. Black-box adversarial test procedure for P0/P1 changes (§3.4) 5. "Break-it" phase for constitutional paths (§3.5) ## Anti-complexity rules (handoff §10, binding) - A concern is not automatically a subsystem. Prefer one invariant / one schema field / one trace relation / one test / one CI gate over a new service. - No future-scale machinery before current-scale evidence requires it. - Every mechanism must answer: trigger / state read / state changed / verification / failure mode / traceability / reversibility / benefit measurement / dormant cost / why an existing authority cannot own it.
conductor added the P1 — high label 2026-09-05 19:22:57 +01:00

⚠️ Priority gate: this issue has no priority. Assign exactly one of P0 — critical (spine/correctness/blocks), P1 — high (material value or real defect), P2 — medium (quality/consolidation), P3 — low (measurement/deferred). An issue without a priority is not filed.

⚠️ **Priority gate:** this issue has no priority. Assign exactly one of **P0 — critical** (spine/correctness/blocks), **P1 — high** (material value or real defect), **P2 — medium** (quality/consolidation), **P3 — low** (measurement/deferred). An issue without a priority is not filed.
gitea-actions Bot added the custodianidentitymemory labels 2026-09-05 19:27:37 +01:00
conductor added P0 — critical and removed P1 — high labels 2026-09-07 11:13:55 +01:00

⚠️ Priority gate: this issue has no priority. Assign exactly one of P0 — critical (spine/correctness/blocks), P1 — high (material value or real defect), P2 — medium (quality/consolidation), P3 — low (measurement/deferred). An issue without a priority is not filed.

⚠️ **Priority gate:** this issue has no priority. Assign exactly one of **P0 — critical** (spine/correctness/blocks), **P1 — high** (material value or real defect), **P2 — medium** (quality/consolidation), **P3 — low** (measurement/deferred). An issue without a priority is not filed.
Author
Owner

Merged via PR #201 (head b77f4b4).

The research round gate is now evidence-enforced:

  • INWARD: ≥2 codebase citations (paths verified)
  • OUTWARD: ≥2 unique external URLs, live-fetched by CI (dead/hallucinated links fail), each with an inline per-source lesson (bare URL lists rejected); forge self-references excluded
  • INNOVATION: must re-reference OUTWARD sources — candidates derive from evidence, not invention
  • SYNTHESIS: evidence-sufficiency statement + realistic feasibility judgement (mechanism/assumptions/failure modes) + kill test (cheapest disproof, carried into the plan as an early frozen test) + plans/ artifact — OR honest MORE-RESEARCH with substantiated open questions (bound: 2 cycles → owner escalation + blocked label)
  • Issue thread = permalinked source-of-truth; rejection comments name exact gaps

Direct response to the observed failure: keyword-only gate cleared an improvised round identically to a genuine one. Verified: YAML valid, logic tested good/bad/self-exclusion, live fetch ok(200) vs unreachable, all four synthesis paths (clear/reject/loopback/loopback-reject) unit-tested.

**Merged via PR #201 (head b77f4b4).** The research round gate is now evidence-enforced: - **INWARD**: ≥2 codebase citations (paths verified) - **OUTWARD**: ≥2 unique external URLs, **live-fetched by CI** (dead/hallucinated links fail), each with an inline per-source lesson (bare URL lists rejected); forge self-references excluded - **INNOVATION**: must re-reference OUTWARD sources — candidates derive from evidence, not invention - **SYNTHESIS**: evidence-sufficiency statement + realistic feasibility judgement (mechanism/assumptions/failure modes) + **kill test** (cheapest disproof, carried into the plan as an early frozen test) + plans/ artifact — OR honest **MORE-RESEARCH** with substantiated open questions (bound: 2 cycles → owner escalation + blocked label) - Issue thread = permalinked source-of-truth; rejection comments name exact gaps Direct response to the observed failure: keyword-only gate cleared an improvised round identically to a genuine one. Verified: YAML valid, logic tested good/bad/self-exclusion, live fetch ok(200) vs unreachable, all four synthesis paths (clear/reject/loopback/loopback-reject) unit-tested.
gitea-actions Bot added the ready label 2026-09-07 16:23:55 +01:00

🎓 NEXT: #147 — Epic: pipeline-integrity hardening tranche (handoff §3) — type gate, coverage ratchet, dep locking, adversarial procedure, break-it phase — dependencies satisfied (scheduled sweep); Rune may start.

🎓 **NEXT: #147 — Epic: pipeline-integrity hardening tranche (handoff §3) — type gate, coverage ratchet, dep locking, adversarial procedure, break-it phase** — dependencies satisfied (scheduled sweep); Rune may start.
Sign in to join this conversation.
2 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: conductor/rune#147