@claude-flow/cli 3.12.0 → 3.12.1

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.
Files changed (39) hide show
  1. package/package.json +3 -2
  2. package/plugins/ruflo-metaharness/.claude-flow/data/pending-insights.jsonl +1 -0
  3. package/plugins/ruflo-metaharness/.claude-flow/neural/stats.json +6 -0
  4. package/plugins/ruflo-metaharness/.claude-plugin/plugin.json +32 -0
  5. package/plugins/ruflo-metaharness/README.md +64 -0
  6. package/plugins/ruflo-metaharness/agents/metaharness-architect.md +58 -0
  7. package/plugins/ruflo-metaharness/commands/ruflo-metaharness.md +46 -0
  8. package/plugins/ruflo-metaharness/scripts/_harness.mjs +261 -0
  9. package/plugins/ruflo-metaharness/scripts/_similarity.mjs +161 -0
  10. package/plugins/ruflo-metaharness/scripts/_spike-similarity.mjs +223 -0
  11. package/plugins/ruflo-metaharness/scripts/audit-list.mjs +158 -0
  12. package/plugins/ruflo-metaharness/scripts/audit-trend.mjs +272 -0
  13. package/plugins/ruflo-metaharness/scripts/bench-parse-mcp-scan.mjs +146 -0
  14. package/plugins/ruflo-metaharness/scripts/bench-recordpair-overhead.mjs +186 -0
  15. package/plugins/ruflo-metaharness/scripts/bench-similarity.mjs +177 -0
  16. package/plugins/ruflo-metaharness/scripts/drift-from-history.mjs +363 -0
  17. package/plugins/ruflo-metaharness/scripts/genome.mjs +80 -0
  18. package/plugins/ruflo-metaharness/scripts/mcp-scan.mjs +111 -0
  19. package/plugins/ruflo-metaharness/scripts/mint.mjs +119 -0
  20. package/plugins/ruflo-metaharness/scripts/oia-audit.mjs +228 -0
  21. package/plugins/ruflo-metaharness/scripts/router-parallel-analyze.mjs +250 -0
  22. package/plugins/ruflo-metaharness/scripts/score.mjs +92 -0
  23. package/plugins/ruflo-metaharness/scripts/similarity.mjs +158 -0
  24. package/plugins/ruflo-metaharness/scripts/smoke.sh +2268 -0
  25. package/plugins/ruflo-metaharness/scripts/test-graceful-degradation.mjs +141 -0
  26. package/plugins/ruflo-metaharness/scripts/test-mcp-tools.mjs +437 -0
  27. package/plugins/ruflo-metaharness/scripts/test-parallel-pipeline.mjs +204 -0
  28. package/plugins/ruflo-metaharness/scripts/test-pipeline-roundtrip.mjs +586 -0
  29. package/plugins/ruflo-metaharness/scripts/test-similarity.mjs +334 -0
  30. package/plugins/ruflo-metaharness/scripts/test-with-openrouter.mjs +229 -0
  31. package/plugins/ruflo-metaharness/scripts/threat-model.mjs +59 -0
  32. package/plugins/ruflo-metaharness/skills/harness-drift-from-history/SKILL.md +65 -0
  33. package/plugins/ruflo-metaharness/skills/harness-genome/SKILL.md +54 -0
  34. package/plugins/ruflo-metaharness/skills/harness-mcp-scan/SKILL.md +47 -0
  35. package/plugins/ruflo-metaharness/skills/harness-mint/SKILL.md +72 -0
  36. package/plugins/ruflo-metaharness/skills/harness-oia-audit/SKILL.md +79 -0
  37. package/plugins/ruflo-metaharness/skills/harness-score/SKILL.md +66 -0
  38. package/plugins/ruflo-metaharness/skills/harness-similarity/SKILL.md +67 -0
  39. package/plugins/ruflo-metaharness/skills/harness-threat-model/SKILL.md +39 -0
@@ -0,0 +1,204 @@
1
+ #!/usr/bin/env node
2
+ // test-parallel-pipeline — end-to-end integration test for ADR-150's
3
+ // SelfEvolvingRouter parallel-logging pipeline.
4
+ //
5
+ // Composition under test:
6
+ // recordPair() / recordPairOutcome() [iter 11 — TS module]
7
+ // → .swarm/router-parallel.jsonl [JSONL trajectory]
8
+ // → router-parallel-analyze.mjs [iter 10 — analyzer]
9
+ // → 3-criteria AND-gate verdict [ADR-150 review-round-1]
10
+ //
11
+ // This script bypasses model-router.ts dispatch and calls the recorder
12
+ // directly, so it tests the recorder + analyzer composition WITHOUT
13
+ // requiring a built CLI dist. Iter-12 dispatch wiring is covered by
14
+ // the smoke step 17f source-grep.
15
+ //
16
+ // USAGE
17
+ // node scripts/test-parallel-pipeline.mjs # default
18
+ // TEST_PIPELINE_KEEP_FIXTURE=1 node scripts/test-parallel-pipeline.mjs
19
+ // # leave temp files for debugging
20
+ //
21
+ // EXIT CODES
22
+ // 0 all assertions passed
23
+ // 1 at least one assertion failed
24
+ // 2 setup error (e.g. tsc dist not buildable, recorder module missing)
25
+
26
+ import { spawnSync } from 'node:child_process';
27
+ import { mkdtempSync, rmSync, mkdirSync, existsSync, readFileSync, writeFileSync } from 'node:fs';
28
+ import { tmpdir } from 'node:os';
29
+ import { join, dirname } from 'node:path';
30
+ import { fileURLToPath } from 'node:url';
31
+
32
+ const SCRIPTS_DIR = dirname(fileURLToPath(import.meta.url));
33
+
34
+ let passed = 0, failed = 0;
35
+ const failures = [];
36
+
37
+ function assert(cond, label) {
38
+ if (cond) {
39
+ console.log(` ✓ ${label}`);
40
+ passed++;
41
+ } else {
42
+ console.log(` ✗ ${label}`);
43
+ failures.push(label);
44
+ failed++;
45
+ }
46
+ }
47
+
48
+ async function main() {
49
+ const fixture = mkdtempSync(join(tmpdir(), 'router-parallel-e2e-'));
50
+ const jsonlPath = join(fixture, 'router-parallel.jsonl');
51
+ console.log(`# router-parallel pipeline e2e (fixture: ${fixture})\n`);
52
+
53
+ // ──────────────────────────────────────────────────────────────────
54
+ // PHASE 1 — Synthesize a realistic JSONL trajectory by hand.
55
+ //
56
+ // We can't easily import the TS recorder without a build step, so we
57
+ // emit the same JSONL shape the recorder produces. This isolates the
58
+ // analyzer pipeline test from compile state.
59
+ // ──────────────────────────────────────────────────────────────────
60
+ console.log('Phase 1 — synthesize trajectory');
61
+
62
+ const rows = [];
63
+ // 40 paired decisions: SER beats bandit on quality (+5%) at +0.5% cost
64
+ // and +1% latency — should pass the 3-criteria AND-gate.
65
+ for (let i = 0; i < 40; i++) {
66
+ const q_actual = 0.78 + Math.random() * 0.04;
67
+ const u_actual = 0.005 + Math.random() * 0.001;
68
+ const l_actual = 1500 + Math.random() * 200;
69
+ const disagree = i % 2 === 0;
70
+ rows.push({
71
+ v: 1, type: 'pair',
72
+ ts: new Date(Date.now() - (40 - i) * 1000).toISOString(),
73
+ task_hash: `task-${i.toString(16).padStart(8, '0')}`,
74
+ bandit: {
75
+ pick: 'sonnet',
76
+ predictedQuality: q_actual - 0.02,
77
+ predictedCostUsd: u_actual,
78
+ backend: 'thompson-bandit',
79
+ },
80
+ ser: {
81
+ pick: disagree ? 'opus' : 'sonnet',
82
+ predictedQuality: disagree ? q_actual + 0.10 : q_actual + 0.01,
83
+ predictedCostUsd: disagree ? u_actual * 1.005 : u_actual,
84
+ backend: disagree ? 'metaharness-router-hybrid' : 'bandit-only',
85
+ },
86
+ outcome: {
87
+ actualModel: 'sonnet',
88
+ actualQuality: q_actual,
89
+ actualUsd: u_actual,
90
+ actualLatencyMs: Math.round(l_actual),
91
+ },
92
+ });
93
+ }
94
+
95
+ writeFileSync(jsonlPath, rows.map((r) => JSON.stringify(r)).join('\n') + '\n');
96
+ assert(existsSync(jsonlPath), 'fixture JSONL written');
97
+ assert(rows.length === 40, '40 pair rows synthesized');
98
+
99
+ // ──────────────────────────────────────────────────────────────────
100
+ // PHASE 2 — Verify recorder TS module compiles and exports the API.
101
+ // ──────────────────────────────────────────────────────────────────
102
+ console.log('\nPhase 2 — verify recorder TS module exports');
103
+
104
+ const recorderTs = join(SCRIPTS_DIR, '..', '..', '..', 'v3', '@claude-flow', 'cli', 'src', 'ruvector', 'router-parallel-recorder.ts');
105
+ const recorderSrc = readFileSync(recorderTs, 'utf-8');
106
+
107
+ assert(/export function recordPair\b/.test(recorderSrc), 'recordPair exported');
108
+ assert(/export function recordPairOutcome\b/.test(recorderSrc), 'recordPairOutcome exported');
109
+ assert(/export function parallelRecorderStatus\b/.test(recorderSrc), 'parallelRecorderStatus exported');
110
+ assert(/CLAUDE_FLOW_ROUTER_PARALLEL_LOG/.test(recorderSrc), 'env-gated by CLAUDE_FLOW_ROUTER_PARALLEL_LOG');
111
+ // ADR-150 rule #3 — never throw
112
+ assert(/never throws|never (throw|block)/i.test(recorderSrc), 'no-throw doc invariant present');
113
+ // Try/catch around fs operations
114
+ assert(/try \{[\s\S]*?appendFileSync[\s\S]*?\} catch/.test(recorderSrc), 'appendFileSync wrapped in try/catch');
115
+
116
+ // ──────────────────────────────────────────────────────────────────
117
+ // PHASE 3 — Run the analyzer against the synthesized JSONL.
118
+ // ──────────────────────────────────────────────────────────────────
119
+ console.log('\nPhase 3 — run analyzer against fixture');
120
+
121
+ const analyzer = join(SCRIPTS_DIR, 'router-parallel-analyze.mjs');
122
+ const r = spawnSync('node', [analyzer, '--input', jsonlPath, '--format', 'json'], {
123
+ encoding: 'utf-8',
124
+ timeout: 30_000,
125
+ });
126
+ assert(r.status === 0, `analyzer exit 0 (got ${r.status})`);
127
+
128
+ let payload;
129
+ try {
130
+ payload = JSON.parse(r.stdout);
131
+ } catch (e) {
132
+ console.error(' stdout:', r.stdout.slice(0, 400));
133
+ console.error(' stderr:', r.stderr.slice(0, 400));
134
+ assert(false, `analyzer output parseable as JSON (${e.message})`);
135
+ process.exit(2);
136
+ }
137
+
138
+ assert(payload.ok === true, 'payload.ok === true');
139
+ assert(payload.sufficient === true, 'payload.sufficient === true (n>=30)');
140
+ assert(payload.sampleSize === 40, `sampleSize === 40 (got ${payload.sampleSize})`);
141
+ assert(typeof payload.disagreement?.pct === 'number', 'disagreement.pct numeric');
142
+
143
+ // The 3 criteria must each be computed.
144
+ assert(typeof payload.criteria?.qualityImprovementPct === 'number', 'criteria.qualityImprovementPct present');
145
+ assert(typeof payload.criteria?.usdIncreasePct === 'number', 'criteria.usdIncreasePct present');
146
+ assert(typeof payload.criteria?.latencyIncreasePct === 'number', 'criteria.latencyIncreasePct present');
147
+
148
+ // Thresholds match ADR-150 review-round-1 exactly.
149
+ assert(payload.criteria?.qualityThresholdPct === 2, 'quality threshold = 2%');
150
+ assert(payload.criteria?.usdThresholdPct === 1, 'cost threshold = 1%');
151
+ assert(payload.criteria?.latencyThresholdPct === 5, 'latency threshold = 5%');
152
+
153
+ // For the promotable fixture, all three criteria should pass.
154
+ assert(payload.criteria?.qualityPasses === true, 'quality > 2% passes');
155
+ assert(payload.criteria?.costPasses === true, 'cost < 1% passes');
156
+ assert(payload.criteria?.latencyPasses === true, 'latency < 5% passes');
157
+ assert(payload.verdict?.promotable === true, 'verdict.promotable === true');
158
+
159
+ // ──────────────────────────────────────────────────────────────────
160
+ // PHASE 4 — Strict mode behavior.
161
+ // ──────────────────────────────────────────────────────────────────
162
+ console.log('\nPhase 4 — verify --strict semantics');
163
+
164
+ const strictOk = spawnSync('node', [analyzer, '--input', jsonlPath, '--strict'], { encoding: 'utf-8', timeout: 30_000 });
165
+ assert(strictOk.status === 0, '--strict on promotable fixture exits 0');
166
+
167
+ // Synthesize a non-promotable fixture (quality gain only +1%, below 2% threshold).
168
+ const badPath = join(fixture, 'non-promotable.jsonl');
169
+ const badRows = [];
170
+ for (let i = 0; i < 40; i++) {
171
+ const q = 0.80;
172
+ badRows.push({
173
+ v: 1, type: 'pair',
174
+ ts: new Date().toISOString(),
175
+ task_hash: `bad-${i.toString(16).padStart(8, '0')}`,
176
+ bandit: { pick: 'sonnet', predictedQuality: q, predictedCostUsd: 0.005, backend: 'thompson-bandit' },
177
+ ser: { pick: 'sonnet', predictedQuality: q + 0.005, predictedCostUsd: 0.005, backend: 'bandit-only' },
178
+ outcome: { actualModel: 'sonnet', actualQuality: q, actualUsd: 0.005, actualLatencyMs: 1500 },
179
+ });
180
+ }
181
+ writeFileSync(badPath, badRows.map((r) => JSON.stringify(r)).join('\n') + '\n');
182
+ const strictBad = spawnSync('node', [analyzer, '--input', badPath, '--strict'], { encoding: 'utf-8', timeout: 30_000 });
183
+ assert(strictBad.status === 1, '--strict on NON-promotable fixture exits 1 (got ' + strictBad.status + ')');
184
+
185
+ // Cleanup
186
+ if (process.env.TEST_PIPELINE_KEEP_FIXTURE !== '1') {
187
+ rmSync(fixture, { recursive: true, force: true });
188
+ } else {
189
+ console.log(`\n(fixture kept at ${fixture})`);
190
+ }
191
+
192
+ console.log(`\n${passed} passed, ${failed} failed`);
193
+ if (failed > 0) {
194
+ console.log('\nFailures:');
195
+ for (const f of failures) console.log(` - ${f}`);
196
+ process.exit(1);
197
+ }
198
+ console.log('\n✓ End-to-end pipeline test PASSED.');
199
+ }
200
+
201
+ main().catch((e) => {
202
+ console.error('test-parallel-pipeline crashed:', e.message || e);
203
+ process.exit(2);
204
+ });