pi-aia-asf 0.12.0 → 0.12.2

package/CHANGELOG.md

@@ -7,6 +7,26 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.12.2] - 2026-10-14
+
+### Added
+
+- (describe changes for 0.12.2)
+
+## [0.12.1] - 2026-10-14
+
+### Added
+
+- **06j §7 — who enforces the model gate.** The boundary is now written down:
+  ASF states the obligation, and the loop guard enforces completion
+  **generically**. When a change touches persistence, the model obligations are
+  captured as MUST specs (`capture_spec`) and marked `met` with evidence; the
+  guard verifies those specs at settle exactly as it verifies any specification
+  arriving from any source. What crosses the boundary is the **specification**,
+  never the check — the guard does not know what a cascade is and must not have
+  to. SKILL.md's Phase 6 model rule now says so directly, and the anti-pattern
+  list names the alternative ("asking the loop guard to understand the model").
+
 ## [0.12.0] - 2026-10-14
 
 ### Added

package/lib/code-health/index.d.mts

@@ -0,0 +1,54 @@
+/**
+ * Type surface for the Code Health gate, which stays plain ESM JavaScript
+ * (lib/code-health/*.mjs) on purpose: it is a build-time harness that shells
+ * out to linters, and rewriting it in TypeScript would buy nothing.
+ *
+ * This file is the boundary — what index.ts is allowed to rely on. It is
+ * hand-maintained, and that is deliberate: a field added to `runHealth`'s
+ * return that is not declared here becomes a type error at the call site
+ * instead of a silent `any`. Before this file, the whole gate was `any` and
+ * `npm run check` could not have noticed.
+ */
+
+export interface HealthRow {
+  key: string;
+  value: number | null;
+  status: string;
+  [extra: string]: unknown;
+}
+
+export interface HealthConfig {
+  enabled: boolean;
+  /** "block" makes violations fail the gate; "warn" reports without blocking. */
+  gate: { mode: "block" | "warn" | string; scope: string };
+  ignore?: string[];
+  [extra: string]: unknown;
+}
+
+export interface HealthResult {
+  /** Human-readable report — what the `/asf health` command prints. */
+  text: string;
+  /** Populated only with `--json`. */
+  json: unknown;
+  /** 1 only when the gate is configured to block and the verdict is unhealthy. */
+  exitCode: 0 | 1;
+  healthy: boolean;
+  rows: HealthRow[];
+  config: HealthConfig;
+}
+
+export interface HealthDeps {
+  readFile?: (path: string) => string | null;
+  tools?: unknown;
+  [extra: string]: unknown;
+}
+
+export declare function runHealth(
+  projectRoot: string,
+  args?: string[],
+  deps?: HealthDeps,
+): Promise<HealthResult>;
+
+export declare function parseArgs(args: string[]): Record<string, boolean | string>;
+
+export declare const BASELINE_FILE: string;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-aia-asf",
-  "version": "0.12.0",
+  "version": "0.12.2",
   "description": "Ai Applied Agentic Software Factory — codifies the full software development flow: intake, research, spec capture, adversarial analysis, planning with approval gates, test-first implementation, and release. Requires pi-vigilant, pi-smart-web-search, pi-smart-fetch, and pi-aia-browser.",
   "keywords": [
     "pi-package",
@@ -31,7 +31,7 @@
   "scripts": {
     "clean": "echo 'nothing to clean'",
     "build": "echo 'nothing to build — pi loads .ts via jiti at runtime'",
-    "check": "tsc --noEmit --target ES2022 --module NodeNext --moduleResolution NodeNext --skipLibCheck --allowImportingTsExtensions --noCheck index.ts",
+    "check": "./scripts/typecheck.sh",
     "test": "./test/run-all.sh",
     "release": "./scripts/release.sh"
   },
@@ -64,10 +64,10 @@
     }
   },
   "devDependencies": {
-    "@earendil-works/pi-agent-core": "0.83.0",
-    "@earendil-works/pi-ai": "0.83.0",
-    "@earendil-works/pi-coding-agent": "0.83.0",
-    "typebox": "1.1.38",
+    "@earendil-works/pi-agent-core": "1.2.0",
+    "@earendil-works/pi-ai": "1.2.0",
+    "@earendil-works/pi-coding-agent": "1.2.0",
+    "typebox": "1.3.27",
     "typescript": "5.9.3"
   }
 }

package/skills/aia-asf/SKILL.md

@@ -236,7 +236,7 @@ Execute the task list milestone by milestone. Discipline rules:
 9. **Challenge approved designs (M4)**: if a spec's literal reading creates product tension (e.g. feedback clusters under "Plan" when "Plan = plans"), stop and resolve it with the user before implementing — never implement blindly and call it delivered.
 10. **Trace before claiming delivered (M5)**: the delivery log is a claim; the code is the evidence. Before marking anything ✅, trace the actual code path, confirm the output is consumed by a surface, and confirm the operator-facing outcome test passes.
 11. **No stubs, no fake data, no invented facts (`references/06i-completeness.md`)**: the deliverable is the real implementation. Finish it, or state the discrepancy plainly to the user — what is missing, why, what remains. Never let the user discover the gap by using the product.
-12. **The model is asserted, not assumed — for any change that WRITES to a store (06b Rule 20 / `references/06j-data-model.md`)**: the trigger is what the change writes, not where the diff lives — "only a backend change" is the case that needs it most. Declare ownership for every relationship touched (composition / instance / ledger, owner, survives-parent), and test the **non-effect**: seed the referencing rows with production-shaped data and assert they survive, and assert the tables the change does not name are unchanged (06g §3). Run `/asf health` and resolve or accept its findings. **Uncertain whether the model is touched? Then it is touched.** A green suite is not evidence here — the table a cascade empties is usually never populated by the suite that stayed green.
+12. **The model is asserted, not assumed — for any change that WRITES to a store (06b Rule 20 / `references/06j-data-model.md`)**: the trigger is what the change writes, not where the diff lives — "only a backend change" is the case that needs it most. Declare ownership for every relationship touched (composition / instance / ledger, owner, survives-parent), and test the **non-effect**: seed the referencing rows with production-shaped data and assert they survive, and assert the tables the change does not name are unchanged (06g §3). Run `/asf health` and resolve or accept its findings. **Uncertain whether the model is touched? Then it is touched.** A green suite is not evidence here — the table a cascade empties is usually never populated by the suite that stayed green. **Capture the obligations as MUST specs (`capture_spec`)** so the loop guard verifies them at settle as it verifies any specification from any source — the guard checks completion, it does not know what a cascade is and must not have to.
 
 ---
 

package/skills/aia-asf/references/06j-data-model.md

@@ -214,6 +214,21 @@ an opinion, and they are the part that works against any store.
       store exposes no constraints.
 - [ ] Or: `n/a: no persistence touched, because <reason>`.
 
+## 7. Who enforces this
+
+ASF states the obligation; the loop guard enforces completion **generically**.
+When a change touches persistence, capture the model obligations as MUST specs
+(`capture_spec`) — the ownership decision per relationship, the invariant and
+non-effect tests, the scan result — and mark each `met` with evidence, not
+assertion. The guard then verifies those specs at settle exactly as it verifies
+any specification that arrives from any source, and refuses "done" while a MUST
+spec is open.
+
+Keep model knowledge on this side of the boundary. The guard does not know what
+a cascade is, does not look at schemas, and must not have to — the day it
+inspects your data model is the day it stops being usable outside this factory.
+What crosses the boundary is the **specification**, never the check.
+
 ## Anti-patterns
 
 - **The instance that inherited its parent's lifecycle** — a fact about a
@@ -237,3 +252,6 @@ an opinion, and they are the part that works against any store.
 - **Silent zero** — a missing filter reported as an empty day. (M8)
 - **Reading a scan exception as a clean deliverable** — an accepted finding is a
   decision you must show, not a pass you may claim. (§4)
+- **Asking the loop guard to understand the model** — it verifies that specs are
+  done; teaching it cascade vocabulary puts a domain check in a general tool.
+  Capture the obligation as a spec instead. (§7)