@surlymochan/evalid 0.5.4 → 0.5.6

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -5,6 +5,22 @@ All notable changes to this project are documented here. Detailed notes live in
5
5
 
6
6
  ## [Unreleased]
7
7
 
8
+ ## [0.5.6] - 2026-11-07
9
+
10
+ ### Changed
11
+ - **Playwright 1.59.1 → 1.65.0 (#49)**: dev-dependency bump; paired Chromium 158.0.8127.18 (build 1254) verified locally. Test evidence: `npm test` 363 tests / 362 pass / 0 fail / 1 skipped (pre-existing); `drift:scan` + `pack:check` green.
12
+ - **js-yaml 4.1.1 → 5.5.0 (major) (#50)**: dependency upgraded; the repo's only production use is `yaml.load` in `src/core/workflow-loader.js` (default schema) plus test-only loads of the two shipped workflows. Migration verified item by item before upgrading (2026-11-07, v4 vs v5 parsers run side by side): both `workflows/yolo.yaml` and `workflows/corps.yaml` parse deep-equal; scalar resolution is unchanged (`on`/`yes` stay strings, `017` stays decimal 17, `0o17` stays octal 15, `1:30` stays a string, anchors/aliases still resolve, duplicate keys still throw); empty input now throws `expected a document, but the input is empty` instead of returning `undefined` (the loader threw either way, the message is better). One real break, handled explicitly: v5's default CORE_SCHEMA no longer resolves the `<<` merge key — a merge key now survives as a literal `"<<"` property, and the loader rejects it with an actionable error ("Inline the merged mapping instead") instead of a confusing schema-validation failure. Pinning `YAML11_SCHEMA` was considered and rejected: it restores additive merging but flips scalars (`on`/`yes` → `true`, `017` → 15, `1:30` → 90) and throws on merge-with-override. The import form also changed with v5 (flat named exports, no default export): the four `import yaml from 'js-yaml'` sites now use `import * as yaml from 'js-yaml'`, call sites unchanged. Pinned by `test/workflow-loader-yaml5.test.js` (7 tests). No bench numbers are involved; the bench-number acceptance for P1-2 / P1-3 / P2-1 / P2-3 still awaits the MiniMax full-matrix batch.
13
+
14
+ ## [0.5.5] - 2026-11-07
15
+
16
+ ### Fixed
17
+ - **Missing package main entry `src/index.js` (#46)**: the published package had no main entry file, so `import '@surlymochan/evalid'` / `require('@surlymochan/evalid')` could not resolve. The entry module is added and pinned by `test/package-entry.test.js`.
18
+
19
+ ### Changed
20
+ - **Dead code removed (#47)**: `writePlanMarkdown` (+ its only caller `buildPlanCommands`), `readPendingGate`, `getLatestMissionSession`, and `resolveAtom` — 51 lines deleted, 0 added. Each was verified dead by repo-wide grep (src, bin, test, docs, scripts, including string-form references): only the definition site remained, with no documentation references. None of the four was part of the public `src/index.js` export surface. Test evidence: full `npm run verify` exit 0 on the merged tree.
21
+
22
+ Acceptance debt, stated plainly: the bench-number acceptance for P1-2 / P1-3 / P2-1 / P2-3 still awaits the MiniMax full-matrix batch (15 instances × 4 arms × reps 3), which is running — no bench numbers are claimed in this release.
23
+
8
24
  ## [0.5.4] - 2026-11-07
9
25
 
10
26
  ### Fixed
package/RELEASE_NOTES.md CHANGED
@@ -1,5 +1,25 @@
1
1
  # Release Notes
2
2
 
3
+ ## 0.5.6 - Playwright 1.65.0; js-yaml 5.5.0
4
+
5
+ Patch release. Two dependency upgrades since 0.5.5, both verified by the full test suite (`npm run verify` exit 0 on the merged tree).
6
+
7
+ **Playwright 1.59.1 → 1.65.0 (#49).** Dev-dependency bump; paired Chromium 158.0.8127.18 (build 1254) verified locally.
8
+
9
+ **js-yaml 4.1.1 → 5.5.0 (#50).** Major upgrade, verified item by item before merging: both shipped workflows parse deep-equal under v4 and v5, scalar resolution is unchanged, and the one real break — v5 no longer resolves the `<<` merge key — is handled explicitly: the workflow loader now rejects a literal `<<` key with an actionable error telling the author to inline the merged mapping, instead of a confusing schema-validation failure. Pinned by `test/workflow-loader-yaml5.test.js` (7 tests).
10
+
11
+ Acceptance debt, stated plainly: the bench-number acceptance for P1-2, P1-3, P2-1 and P2-3 — the wall-clock/token/resolved lines from issue #5 — still awaits the MiniMax full-matrix batch (15 instances × 4 arms × reps 3), which is running. No bench numbers are claimed in this release.
12
+
13
+ ## 0.5.5 - Package entry restored; dead code removed
14
+
15
+ Patch release. Two small fixes since 0.5.4, both verified by the full test suite (`npm run verify` exit 0 on the merged tree).
16
+
17
+ **Missing package main entry (#46).** The published package had no main entry file, so importing the package by name could not resolve. `src/index.js` is now present as the package entry and is pinned by `test/package-entry.test.js`.
18
+
19
+ **Dead code removed (#47).** Four exported functions with zero callers are deleted (51 lines, 0 added): `writePlanMarkdown` (plus `buildPlanCommands`, which only it called), `readPendingGate`, `getLatestMissionSession`, and `resolveAtom`. Each was verified dead by repo-wide grep across src, bin, test, docs, and scripts, including string-form references — only the definition site remained in every case. None of the four was exported from the public `src/index.js` surface, so this is not a public API change. The modules themselves stay: their other exports (`load` / `dryRun`, `listPendingGates` / `ackPendingGate`, `getLatestPhaseRun`, `writeReviewArtifact`) are still used.
20
+
21
+ Acceptance debt, stated plainly: the bench-number acceptance for P1-2, P1-3, P2-1 and P2-3 — the wall-clock/token/resolved lines from issue #5 — still awaits the MiniMax full-matrix batch (15 instances × 4 arms × reps 3), which is running. No bench numbers are claimed in this release.
22
+
3
23
  ## 0.5.4 - Phase-cost drift test stops flaking under load
4
24
 
5
25
  Patch release. One test-only fix: `npm run verify` was red on main because the phase-cost drift test tripped its 5% bound at 5.7% while the MiniMax batch was running, then went 7/7 green on a quiet re-run of the same file.
@@ -1,6 +1,6 @@
1
1
  # evalid Launch Dossier
2
2
 
3
- Generated: 2026-11-07T07:29:35.000Z
3
+ Generated: 2026-11-07T15:22:52.000Z
4
4
  Mode: post_publish
5
5
  Verdict: conditionally ready: engine and adoption evidence are strong, registry publication remains required for a splash launch
6
6
  Launch ready: not yet
@@ -117,7 +117,7 @@ Evidence:
117
117
  - Adoption records: 1; status: pass
118
118
  - Goal audit: pass; goal: Keep evalid credible as an open-source agent delivery runtime: repo-owned workflow contracts, executable verification gates, durable project context, and honest benchmark evidence.
119
119
  - Package evidence: missing; source: registry_json
120
- - Package name/version: @surlymochan/evalid@0.5.4
120
+ - Package name/version: @surlymochan/evalid@0.5.6
121
121
 
122
122
  ## Passing Evidence
123
123
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@surlymochan/evalid",
3
- "version": "0.5.4",
3
+ "version": "0.5.6",
4
4
  "description": "Claimed != verified: a runtime verification layer for coding agents. It seals each test run and replays it from a clean snapshot, with a zero-judge gate pipeline (spec clarification, code review, TDD proof) and the xflow CLI.",
5
5
  "keywords": [
6
6
  "ai-agents",
@@ -77,9 +77,9 @@
77
77
  },
78
78
  "dependencies": {
79
79
  "ajv": "^8.12.0",
80
- "js-yaml": "^4.1.0"
80
+ "js-yaml": "^5.5.0"
81
81
  },
82
82
  "devDependencies": {
83
- "playwright": "^1.59.1"
83
+ "playwright": "^1.65.0"
84
84
  }
85
85
  }
@@ -271,32 +271,6 @@ export async function synthesizePhaseArtifact(input = {}, context = {}, options
271
271
  };
272
272
  }
273
273
 
274
- export function buildPlanCommands(projectRoot) {
275
- const packageJsonPath = resolve(projectRoot, 'package.json');
276
- const packageJson = readJsonFile(packageJsonPath, null);
277
- if (packageJson?.scripts?.test) {
278
- return ['npm test'];
279
- }
280
- return ['echo "no test script configured"'];
281
- }
282
-
283
- export function writePlanMarkdown(projectRoot, changeId, phase = 'plan') {
284
- const commands = buildPlanCommands(projectRoot);
285
- const mdPath = resolveChangePath(projectRoot, changeId, `${phase}.md`);
286
- const lines = [
287
- '# Plan',
288
- '',
289
- '## Verification',
290
- '',
291
- '```bash',
292
- ...commands,
293
- '```',
294
- '',
295
- ];
296
- writeTextFile(mdPath, lines.join('\n'));
297
- return mdPath;
298
- }
299
-
300
274
  export function writeReviewArtifact(projectRoot, changeId, summary, findings = []) {
301
275
  const outputPath = resolveChangePath(projectRoot, changeId, 'review.json');
302
276
  writeJsonFile(outputPath, {
@@ -34,15 +34,6 @@ export function listPendingGates(projectRoot = process.cwd()) {
34
34
  .sort((left, right) => String(right.created_at || '').localeCompare(String(left.created_at || '')));
35
35
  }
36
36
 
37
- export function readPendingGate(projectRoot = process.cwd(), phase) {
38
- const gateFile = resolve(resolvePendingGatesDir(projectRoot), `${phase}.json`);
39
- if (!existsSync(gateFile)) {
40
- return null;
41
- }
42
- const gate = readGateFile(gateFile);
43
- return gate ? { ...gate, gate_file: gateFile } : null;
44
- }
45
-
46
37
  export function ackPendingGate(projectRoot = process.cwd(), phase, options = {}) {
47
38
  const gateFile = resolve(resolvePendingGatesDir(projectRoot), `${phase}.json`);
48
39
  if (!existsSync(gateFile)) {
@@ -1,5 +1,3 @@
1
- import { listPhaseRunSessions } from './mission-state.js';
2
-
3
1
  function ensureArray(value) {
4
2
  return Array.isArray(value) ? value : [];
5
3
  }
@@ -21,7 +19,3 @@ export function getLatestPhaseEntry(phaseRun, phase) {
21
19
  const entries = ensureArray(phaseRun.phases).filter((entry) => entry.phase === phase);
22
20
  return entries.sort((left, right) => String(right.updated_at || '').localeCompare(String(left.updated_at || '')))[0] || null;
23
21
  }
24
-
25
- export function getLatestMissionSession(state, missionId) {
26
- return listPhaseRunSessions(state, missionId)[0] || null;
27
- }
@@ -16,7 +16,7 @@
16
16
  import { readFileSync, existsSync } from 'fs';
17
17
  import { resolve, dirname } from 'path';
18
18
  import { fileURLToPath } from 'url';
19
- import yaml from 'js-yaml';
19
+ import * as yaml from 'js-yaml';
20
20
  import Ajv from 'ajv';
21
21
 
22
22
  const __dirname = dirname(fileURLToPath(import.meta.url));
@@ -54,6 +54,13 @@ export function load(workflowPath, { skipRuntimeChecks = false } = {}) {
54
54
  const raw = readFileSync(absPath, 'utf8');
55
55
  const workflow = yaml.load(raw);
56
56
 
57
+ // js-yaml 5 loads with CORE_SCHEMA, which no longer resolves the `<<`
58
+ // merge key: a merge key now survives parsing as a literal "<<" property
59
+ // instead of merging. The schema below (additionalProperties: false)
60
+ // would reject it with a confusing validation error, so fail here with
61
+ // an actionable one. Anchors/aliases still work; only `<<` is affected.
62
+ assertNoMergeKeys(workflow, absPath);
63
+
57
64
  // 1. Schema validation
58
65
  const schema = loadSchema();
59
66
  const ajv = new Ajv({ allErrors: true });
@@ -108,6 +115,29 @@ export function load(workflowPath, { skipRuntimeChecks = false } = {}) {
108
115
 
109
116
  // ─── Runtime checks ──────────────────────────────────────────────────────────
110
117
 
118
+ /**
119
+ * Reject literal `<<` keys left behind by js-yaml 5's CORE_SCHEMA default,
120
+ * which no longer resolves YAML merge keys. A workflow that relied on
121
+ * merging must inline the merged mapping instead. Walks the whole parsed
122
+ * document so the error names the exact location.
123
+ */
124
+ function assertNoMergeKeys(value, workflowPath, trail = '$') {
125
+ if (Array.isArray(value)) {
126
+ value.forEach((item, i) => assertNoMergeKeys(item, workflowPath, `${trail}[${i}]`));
127
+ return;
128
+ }
129
+ if (value && typeof value === 'object') {
130
+ for (const [key, child] of Object.entries(value)) {
131
+ if (key === '<<') {
132
+ throw new Error(
133
+ `Workflow uses a YAML merge key ("<<") at ${trail} in ${workflowPath}, which js-yaml 5 no longer resolves. Inline the merged mapping instead.`
134
+ );
135
+ }
136
+ assertNoMergeKeys(child, workflowPath, `${trail}.${key}`);
137
+ }
138
+ }
139
+ }
140
+
111
141
  import { spawnSync } from 'child_process';
112
142
 
113
143
  function checkBinOnPath(bin, description) {
@@ -151,13 +181,3 @@ export function dryRun(workflowPath, input = {}) {
151
181
 
152
182
  return { workflow, phases };
153
183
  }
154
-
155
- /**
156
- * Resolve a single atom from the registry to its invocation descriptor.
157
- * Used by the execution layer to dispatch atoms.
158
- */
159
- export function resolveAtom(atomId, registry) {
160
- const def = registry.atoms[atomId];
161
- if (!def) throw new Error(`Unknown atom: ${atomId}`);
162
- return { ...def, id: atomId };
163
- }
package/src/index.js ADDED
@@ -0,0 +1,12 @@
1
+ /**
2
+ * Programmatic entry point for @surlymochan/evalid.
3
+ *
4
+ * This is the file package.json "main" points to. It used to not exist at
5
+ * all, so `import '@surlymochan/evalid'` / `require('@surlymochan/evalid')`
6
+ * failed with MODULE_NOT_FOUND even though the published tarball shipped
7
+ * "main": "src/index.js". The package is CLI-first (bin: xflow), but the
8
+ * control-plane server is the one documented programmatic surface, so this
9
+ * module re-exports its starter. Importing this module has no side effects:
10
+ * it does not start a server.
11
+ */
12
+ export { default, default as startServer } from './server.js';