instar 1.3.470 → 1.3.471

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "instar",
-  "version": "1.3.470",
+  "version": "1.3.471",
   "description": "Coherence infrastructure for self-evolving AI agents — on the Claude Code or Codex subscription you already have.",
   "type": "module",
   "main": "dist/index.js",

package/skills/spec-converge/SKILL.md

@@ -69,7 +69,7 @@ The skill spawns reviewers in parallel:
 - **Scalability/performance.** Hot-path cost, concurrent writes, memory churn, fail-open semantics, hook latency.
 - **Adversarial.** Misbehaving-session scenarios — bad-entry poisoning, self-reinforcing loops, stale claims, authority ambiguity, kind gaming.
 - **Integration/deployment.** Migration, backup/restore, multi-machine, config knobs, dashboard surface, rollback.
-- **Lessons-aware.** Loads the canonical Instar Design Principles + Lessons Learned index (`docs/INSTAR-DESIGN-PRINCIPLES-AND-LESSONS.md`) plus the running agent's local `.instar/memory/feedback_*.md` entries, then checks the spec for (a) direct contradictions of documented principles/lessons, (b) applicable lessons the spec fails to engage with, and (c) behavioral lessons violated by agent-facing surfaces the spec proposes. Catches the "Phase 2" anti-pattern and the spec-converge-pre-auth-circular failure mode (see `feedback_spec_converge_pre_auth_circular`).
+- **Lessons-aware.** Loads the canonical Instar Design Principles + Lessons Learned index (`docs/INSTAR-DESIGN-PRINCIPLES-AND-LESSONS.md`) plus the running agent's local `.instar/memory/feedback_*.md` entries, then checks the spec for (a) direct contradictions of documented principles/lessons, (b) applicable lessons the spec fails to engage with, (c) behavioral lessons violated by agent-facing surfaces the spec proposes, and **(d) FOUNDATION/SUBSYSTEM AUDIT — the design the spec TESTS, EXTENDS, or BUILDS ON, not just the spec's own surface: does that foundation violate a known standard or repeat a known mistake?** The audit MUST reach one layer below the spec boundary. A spec can be internally clean while faithfully testing or extending a flawed foundation — e.g. a test-harness spec that correctly proves a permission gate which *itself* holds brittle blocking authority in violation of Signal-vs-Authority, or an extension spec built on a subsystem with an unaddressed gap. Taking the underlying subsystem "as given" is exactly how a standards violation survives review: the spec passes, the foundation's flaw is never weighed. When the foundation is flawed, the finding is "this spec is sound but the subsystem it depends on violates standard X / repeats mistake Y — surface it before building on/around it." Catches the "Phase 2" anti-pattern, the spec-converge-pre-auth-circular failure mode (see `feedback_spec_converge_pre_auth_circular`), and the foundation-not-audited gap (the Slack-permission-gate brittle-enum-as-authority that the harness convergence cleared because it audited only the harness, 2026-06-09).
 
 **External reviewer (cross-model, via the agent's own installed codex CLI):**
 

package/src/data/builtin-manifest.json

@@ -1,8 +1,8 @@
 {
   "$schema": "./builtin-manifest.schema.json",
   "schemaVersion": 1,
-  "generatedAt": "2026-06-10T16:53:50.243Z",
-  "instarVersion": "1.3.470",
+  "generatedAt": "2026-06-10T18:02:55.574Z",
+  "instarVersion": "1.3.471",
   "entryCount": 201,
   "entries": {
     "hook:session-start": {
@@ -11,7 +11,7 @@
       "domain": "identity",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/session-start.sh",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:dangerous-command-guard": {
@@ -20,7 +20,7 @@
       "domain": "safety",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/dangerous-command-guard.sh",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:grounding-before-messaging": {
@@ -29,7 +29,7 @@
       "domain": "safety",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/grounding-before-messaging.sh",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:compaction-recovery": {
@@ -38,7 +38,7 @@
       "domain": "identity",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/compaction-recovery.sh",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:external-operation-gate": {
@@ -47,7 +47,7 @@
       "domain": "safety",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/external-operation-gate.js",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:deferral-detector": {
@@ -56,7 +56,7 @@
       "domain": "safety",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/deferral-detector.js",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:self-stop-guard": {
@@ -65,7 +65,7 @@
       "domain": "coherence",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/self-stop-guard.js",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:post-action-reflection": {
@@ -74,7 +74,7 @@
       "domain": "evolution",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/post-action-reflection.js",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:external-communication-guard": {
@@ -83,7 +83,7 @@
       "domain": "safety",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/external-communication-guard.js",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:scope-coherence-collector": {
@@ -92,7 +92,7 @@
       "domain": "coherence",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/scope-coherence-collector.js",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:scope-coherence-checkpoint": {
@@ -101,7 +101,7 @@
       "domain": "coherence",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/scope-coherence-checkpoint.js",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:free-text-guard": {
@@ -110,7 +110,7 @@
       "domain": "safety",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/free-text-guard.sh",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:claim-intercept": {
@@ -119,7 +119,7 @@
       "domain": "coherence",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/claim-intercept.js",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:claim-intercept-response": {
@@ -128,7 +128,7 @@
       "domain": "coherence",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/claim-intercept-response.js",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:stop-gate-router": {
@@ -137,7 +137,7 @@
       "domain": "safety",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/stop-gate-router.js",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "hook:auto-approve-permissions": {
@@ -146,7 +146,7 @@
       "domain": "safety",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
       "installedPath": ".instar/hooks/instar/auto-approve-permissions.js",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "job:health-check": {
@@ -1554,7 +1554,7 @@
       "type": "subsystem",
       "domain": "updates",
       "sourcePath": "src/core/PostUpdateMigrator.ts",
-      "contentHash": "c4b6d22bcb02a0ce956d38326c9d764b47bac8c37545ccf287bef29f9912f908",
+      "contentHash": "91a08551e1f1c321ae5b923cc68577dd8bda1381bc810d8e3444295947b00ab4",
       "since": "2025-01-01"
     },
     "subsystem:scheduler": {

package/upgrades/1.3.471.md

@@ -0,0 +1,50 @@
+# Upgrade Guide — vNEXT
+
+<!-- assembled-by: assemble-next-md -->
+<!-- bump: patch -->
+
+## What Changed
+
+The `/spec-converge` review's Lessons-aware reviewer gained a new mandate — clause
+**(d) FOUNDATION/SUBSYSTEM AUDIT**: the review must now reach one layer below the spec
+boundary and weigh the subsystem the spec *tests, extends, or builds on* against known
+standards and lessons, not just the spec's own surface. When that foundation is flawed,
+the finding is "this spec is sound but the subsystem it depends on violates standard X /
+repeats mistake Y — surface it before building on/around it."
+
+This closes a real review gap (2026-06-09): a test-harness spec converged cleanly while
+the permission gate it proved still held brittle blocking authority in violation of
+Signal-vs-Authority. The convergence audited only the harness and took the flawed
+foundation as given. A new idempotent `PostUpdateMigrator` step
+(`migrateSpecConvergeFoundationAudit`) backfills the clause into existing agents'
+installed spec-converge skill on update (fingerprint-guarded; customized skills are left
+untouched).
+
+## What to Tell Your User
+
+Nothing changes for end users — this is an internal hardening of the instar
+spec-review process used by instar-developing agents. No runtime behavior, gate, or
+message path changes. The practical effect is that future instar specs are reviewed one
+layer deeper, so a spec can no longer pass review while quietly testing or extending a
+component that itself violates a standard.
+
+## Summary of New Capabilities
+
+- `/spec-converge` Lessons-aware reviewer clause (d): foundation/subsystem audit one
+  layer below the spec boundary.
+- `PostUpdateMigrator.migrateSpecConvergeFoundationAudit`: idempotent, fingerprint-guarded
+  backfill of the clause for existing agents (customized skills untouched).
+
+## Evidence
+
+- `skills/spec-converge/SKILL.md` — clause (d) added to the Lessons-aware bullet.
+- `src/core/PostUpdateMigrator.ts` — `migrateSpecConvergeFoundationAudit` registered in
+  the migrate runner; mirrors the established fingerprint-guarded skill-content migration
+  idiom (`migrateInstarDevInternalOnlyReleaseNoteLane`, `migrateTestAsSelfSkill`).
+- `tests/unit/migrate-spec-converge-foundation-audit.test.ts` — 5 tests: updates a stock
+  skill, idempotent when the marker is present, leaves a customized skill untouched,
+  no-ops when not installed, and a wiring-integrity test asserting the bundled source
+  actually carries the marker (so the migration cannot silently no-op for every agent).
+- `tsc --noEmit` clean; sibling skill-migration + builtin-dev-skills suites green (19
+  tests across the touched area).
+- Side-effects review: `upgrades/side-effects/spec-converge-foundation-audit.md`.

package/upgrades/side-effects/spec-converge-foundation-audit.md

@@ -0,0 +1,92 @@
+# Side-Effects Review — spec-converge clause (d) foundation/subsystem audit
+
+**Version / slug:** `spec-converge-foundation-audit`
+**Date:** `2026-06-10`
+**Author:** `echo`
+**Tier:** `1` (skill-prompt clause + idempotent content migration + tests; no new runtime subsystem, route, or config)
+**Second-pass reviewer:** `not required`
+
+## Summary of the change
+
+Extend the `/spec-converge` Lessons-aware reviewer's mandate with clause **(d)
+FOUNDATION/SUBSYSTEM AUDIT**: the review must reach one layer below the spec boundary
+and weigh the subsystem the spec *tests / extends / builds on* against known standards
+and lessons — not just the spec's own surface. When that foundation is flawed, the
+finding is "this spec is sound but the subsystem it depends on violates standard X /
+repeats mistake Y — surface it before building on/around it."
+
+Two edits + tests:
+1. `skills/spec-converge/SKILL.md` — the bundled source skill gains clause (d) on the
+   Lessons-aware bullet (the `## Internal reviewers` section).
+2. `src/core/PostUpdateMigrator.ts` — a new `migrateSpecConvergeFoundationAudit`
+   (registered in the migrate runner) so existing agents' installed
+   `.claude/skills/spec-converge/SKILL.md` receives clause (d) on update. Idempotent +
+   conservative: re-copies the bundled skill only when the installed copy (a) lacks the
+   `FOUNDATION/SUBSYSTEM AUDIT` marker AND (b) still matches the stock spec-converge
+   fingerprint (`# /spec-converge` + `**Internal reviewers (Claude subagents):**`). A
+   customized skill is skipped untouched.
+3. `tests/unit/migrate-spec-converge-foundation-audit.test.ts` — 5 tests (update,
+   idempotent, customized-untouched, not-installed no-op, bundled-source wiring
+   integrity).
+
+Motivation: the 2026-06-09 gap where a test-harness spec converged cleanly while the
+permission gate it proved still held brittle blocking authority in violation of
+Signal-vs-Authority. The convergence audited only the harness and took the flawed
+foundation as given. This closes that gap structurally (Structure > Willpower): the
+reviewer is *instructed* to look below the spec, rather than the author having to
+remember to.
+
+## Decision-point inventory
+
+- **Migration content-sniff guard** — re-copy vs. skip. Inputs: installed skill text.
+  Branch 1: marker already present → return (idempotent no-op). Branch 2: stock
+  fingerprint missing → skip + record `customized`. Branch 3: stock fingerprint present
+  AND marker absent → re-copy the bundled skill. No LLM, no runtime authority — pure
+  string-presence checks at update time.
+
+## 1. Over-block
+
+**What legitimate inputs does this change reject?** Nothing is rejected at runtime — no
+gate, no message path, no API. The migration's only "rejection" is *declining to update*
+a skill it can't confidently recognize as stock (fingerprint missing → skipped). That is
+the safe direction: a customized spec-converge skill keeps the operator's edits rather
+than being clobbered. The reviewer clause makes the review *stricter* (one more thing to
+check), but the Lessons-aware reviewer is signal-only — it surfaces findings, it does not
+block; blocking authority remains the separate Standards-Conformance Gate path.
+
+## 2. Under-block
+
+**What does this still miss?** (a) The clause is a *prompt* instruction, so its
+effectiveness depends on the reviewing model actually reaching one layer down — it is a
+stronger signal, not a code-enforced guarantee (a future code-backed "foundation
+referenced by the spec was audited" check could harden it, but that is out of scope for a
+Tier-1 prompt change). (b) The migration re-copies the *whole* bundled skill (matching the
+established skill-migration idiom); an agent on a much older stock version receives all
+accumulated bundled changes, not just clause (d) — acceptable because the fingerprint
+guard only fires on stock copies. (c) Agents who customized the skill are intentionally
+not upgraded; they must adopt clause (d) by hand.
+
+## 3. Level-of-abstraction fit
+
+**Right layer?** Yes. The clause lives in the one reviewer whose job is standards/lessons
+conformance (Lessons-aware), beside its existing (a)/(b)/(c) checks — not bolted onto an
+unrelated reviewer. The migration lives beside the sibling skill-content migrations
+(`migrateInstarDevInternalOnlyReleaseNoteLane`, `migrateTestAsSelfSkill`) and reuses their
+exact fingerprint-guarded re-copy pattern, so it inherits the proven Migration-Parity
+idiom rather than inventing a new one.
+
+## Migration / rollback
+
+- **Migration:** `migrateSpecConvergeFoundationAudit` runs on the normal post-update path.
+  Idempotent (marker check) and conservative (fingerprint guard). No state, no config.
+- **Rollback:** pure revert — drop the clause from the bundled skill, remove the migration
+  method + its registration line, delete the test. No data migration to undo; the migration
+  only ever copied a newer skill prompt over a stock older one.
+
+## Testing-integrity note
+
+Unit tests cover both sides of the migration's decision boundary (updates stock /
+skips customized / idempotent / not-installed) plus a wiring-integrity test asserting the
+bundled source actually carries the marker (so the migration cannot silently no-op for
+every agent). `tsc --noEmit` clean; sibling skill-migration + builtin-dev-skills suites
+stay green (19 tests total across the touched area).