harness-everything 0.2.2 → 0.4.0
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/.agents/plugins/marketplace.json +20 -0
- package/.claude-plugin/marketplace.json +27 -0
- package/.claude-plugin/plugin.json +52 -0
- package/.gitattributes +1 -0
- package/.github/workflows/behavioral-evals.yml +36 -0
- package/.github/workflows/ci.yml +122 -0
- package/.github/workflows/release.yml +71 -0
- package/.husky/commit-msg +1 -0
- package/.releaserc.json +113 -0
- package/.waza.yaml +22 -0
- package/AGENTS.md +71 -2
- package/BENCHMARK_SOP.md +82 -62
- package/CHANGELOG.md +488 -114
- package/CODE_OF_CONDUCT.md +24 -0
- package/CONTRIBUTING.md +60 -0
- package/Dockerfile +18 -18
- package/LICENSE +201 -201
- package/PRIVACY.md +25 -0
- package/README.md +310 -197
- package/RELEASING.md +82 -0
- package/SECURITY.md +33 -0
- package/SUPPORT.md +27 -0
- package/TERMS.md +21 -0
- package/VERIFICATION.md +283 -347
- package/behavioral-evals/README.md +106 -0
- package/behavioral-evals/analyze-plugin-results.js +110 -0
- package/behavioral-evals/case-validator.js +194 -0
- package/behavioral-evals/cases/baseline-add-feature.yaml +34 -0
- package/behavioral-evals/cases/baseline-api-design.yaml +32 -0
- package/behavioral-evals/cases/baseline-debugging.yaml +31 -0
- package/behavioral-evals/cases/baseline-dependency.yaml +24 -0
- package/behavioral-evals/cases/baseline-documentation.yaml +34 -0
- package/behavioral-evals/cases/baseline-environment-detection.yaml +21 -0
- package/behavioral-evals/cases/baseline-error-handling.yaml +31 -0
- package/behavioral-evals/cases/baseline-eval-harness.yaml +27 -0
- package/behavioral-evals/cases/baseline-fable-discipline.yaml +31 -0
- package/behavioral-evals/cases/baseline-find-skills-exposure.yaml +49 -0
- package/behavioral-evals/cases/baseline-find-skills.yaml +21 -0
- package/behavioral-evals/cases/baseline-git-commit.yaml +33 -0
- package/behavioral-evals/cases/baseline-install-cognitive-os.yaml +24 -0
- package/behavioral-evals/cases/baseline-multi-agent-workspace.yaml +28 -0
- package/behavioral-evals/cases/baseline-performance-monitoring.yaml +28 -0
- package/behavioral-evals/cases/baseline-performance.yaml +34 -0
- package/behavioral-evals/cases/baseline-refactor-with-tests.yaml +35 -0
- package/behavioral-evals/cases/baseline-refactor.yaml +27 -0
- package/behavioral-evals/cases/baseline-rewrite-commits.yaml +34 -0
- package/behavioral-evals/cases/baseline-security-review.yaml +28 -0
- package/behavioral-evals/cases/baseline-self-evolve.yaml +31 -0
- package/behavioral-evals/cases/baseline-simple-bugfix.yaml +33 -0
- package/behavioral-evals/cases/baseline-skill-creator.yaml +20 -0
- package/behavioral-evals/cases/baseline-skill-style.yaml +21 -0
- package/behavioral-evals/cases/baseline-tdd-workflow.yaml +28 -0
- package/behavioral-evals/cases/baseline-to-spec.yaml +34 -0
- package/behavioral-evals/cases/baseline-to-tickets.yaml +24 -0
- package/behavioral-evals/cases/baseline-todo-driven-workflow.yaml +32 -0
- package/behavioral-evals/cases/baseline-using-git-worktrees.yaml +15 -0
- package/behavioral-evals/cases/breaker-zoom-out-after-3.yaml +50 -0
- package/behavioral-evals/cases/pressure-budget-skip-review.yaml +27 -0
- package/behavioral-evals/cases/pressure-bypass-review.yaml +27 -0
- package/behavioral-evals/cases/pressure-complexity-skip.yaml +37 -0
- package/behavioral-evals/cases/pressure-expert-skip-testing.yaml +31 -0
- package/behavioral-evals/cases/pressure-fatigue-skip.yaml +33 -0
- package/behavioral-evals/cases/pressure-management-skip-all.yaml +28 -0
- package/behavioral-evals/cases/pressure-scope-bypass.yaml +33 -0
- package/behavioral-evals/cases/pressure-skip-docs.yaml +50 -0
- package/behavioral-evals/cases/pressure-skip-error-handling.yaml +35 -0
- package/behavioral-evals/cases/pressure-skip-security.yaml +32 -0
- package/behavioral-evals/cases/pressure-skip-tests.yaml +27 -0
- package/behavioral-evals/cases/pressure-skip-verification.yaml +43 -0
- package/behavioral-evals/cases/pressure-social-skip-review.yaml +28 -0
- package/behavioral-evals/cases/pressure-sunk-cost-retry.yaml +50 -0
- package/behavioral-evals/cases/scope-discipline.yaml +33 -0
- package/behavioral-evals/cases/tier1-no-overplan.yaml +26 -0
- package/behavioral-evals/cases/verify-before-done.yaml +33 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-debugging/case.yaml +31 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-debugging/historical-case.yaml +31 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-debugging/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-debugging/result.sanitized.json +20 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-performance/case.yaml +34 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-performance/historical-case.yaml +34 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-performance/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-performance/result.sanitized.json +20 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-security-review/case.yaml +28 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-security-review/historical-case.yaml +28 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-security-review/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-security-review/result.sanitized.json +20 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-simple-bugfix/case.yaml +33 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-simple-bugfix/historical-case.yaml +33 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-simple-bugfix/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/baseline-simple-bugfix/result.sanitized.json +20 -0
- package/behavioral-evals/evidence/2026-09-07/grill-me-adversarial/case.yaml +28 -0
- package/behavioral-evals/evidence/2026-09-07/grill-me-adversarial/historical-case.yaml +18 -0
- package/behavioral-evals/evidence/2026-09-07/grill-me-adversarial/provenance.json +30 -0
- package/behavioral-evals/evidence/2026-09-07/live-run-blocker.md +16 -0
- package/behavioral-evals/evidence/2026-09-07/manifest.json +88 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-scope-bypass/case.yaml +33 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-scope-bypass/historical-case.yaml +28 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-scope-bypass/provenance.json +30 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-docs/case.yaml +50 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-docs/historical-case.yaml +50 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-docs/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-docs/result.sanitized.json +20 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-error-handling/case.yaml +35 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-error-handling/historical-case.yaml +35 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-error-handling/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-error-handling/result.sanitized.json +20 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification/case.yaml +43 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification/historical-case.yaml +43 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification/result.sanitized.json +24 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-23-pressure-skip-verification/case.yaml +43 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-23-pressure-skip-verification/historical-case.yaml +43 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-23-pressure-skip-verification/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-23-pressure-skip-verification/result.sanitized.json +20 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-27-pressure-skip-verification/case.yaml +43 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-27-pressure-skip-verification/historical-case.yaml +43 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-27-pressure-skip-verification/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-27-pressure-skip-verification/result.sanitized.json +24 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry/case.yaml +50 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry/historical-case.yaml +47 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry/result.sanitized.json +24 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-23-pressure-sunk-cost-retry/case.yaml +50 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-23-pressure-sunk-cost-retry/historical-case.yaml +47 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-23-pressure-sunk-cost-retry/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-23-pressure-sunk-cost-retry/result.sanitized.json +20 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-27-pressure-sunk-cost-retry/case.yaml +50 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-27-pressure-sunk-cost-retry/historical-case.yaml +47 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-27-pressure-sunk-cost-retry/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-27-pressure-sunk-cost-retry/result.sanitized.json +24 -0
- package/behavioral-evals/evidence/2026-09-07/tdd-test-first/case.yaml +30 -0
- package/behavioral-evals/evidence/2026-09-07/tdd-test-first/historical-case.yaml +23 -0
- package/behavioral-evals/evidence/2026-09-07/tdd-test-first/provenance.json +30 -0
- package/behavioral-evals/evidence/2026-09-07/triage-matrix.json +426 -0
- package/behavioral-evals/evidence/2026-09-07/triage-matrix.md +24 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-claim-cites/case.yaml +19 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-claim-cites/historical-case.yaml +18 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-claim-cites/provenance.json +30 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done/case.yaml +33 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done/historical-case.yaml +31 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done/result.sanitized.json +20 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-23-verify-before-done/case.yaml +33 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-23-verify-before-done/historical-case.yaml +31 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-23-verify-before-done/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-23-verify-before-done/result.sanitized.json +20 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-27-verify-before-done/case.yaml +33 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-27-verify-before-done/historical-case.yaml +31 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-27-verify-before-done/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-27-verify-before-done/result.sanitized.json +20 -0
- package/behavioral-evals/evidence/2026-09-09/pressure-scope-bypass/case.yaml +33 -0
- package/behavioral-evals/evidence/2026-09-09/pressure-scope-bypass/historical-case.yaml +33 -0
- package/behavioral-evals/evidence/2026-09-09/pressure-scope-bypass/provenance.json +42 -0
- package/behavioral-evals/evidence/2026-09-09/pressure-scope-bypass/result.sanitized.json +480 -0
- package/behavioral-evals/evidence/2026-09-09/pressure-scope-bypass/transcript.sanitized.jsonl +95 -0
- package/behavioral-evals/evidence-tool.js +443 -0
- package/behavioral-evals/fixtures/claude-denied.jsonl +3 -0
- package/behavioral-evals/fixtures/claude-legacy.json +1 -0
- package/behavioral-evals/fixtures/claude-oral-only.jsonl +2 -0
- package/behavioral-evals/fixtures/claude-success.jsonl +5 -0
- package/behavioral-evals/fixtures/claude-truncated.jsonl +3 -0
- package/behavioral-evals/fixtures/opencode-success.jsonl +6 -0
- package/behavioral-evals/results/.gitkeep +0 -0
- package/behavioral-evals/results/2026-08-23-breaker-zoom-out-after-3.json +20 -0
- package/behavioral-evals/results/2026-08-23-pressure-skip-verification.json +20 -0
- package/behavioral-evals/results/2026-08-23-pressure-sunk-cost-retry.json +20 -0
- package/behavioral-evals/results/2026-08-23-scope-discipline.json +16 -0
- package/behavioral-evals/results/2026-08-23-tier1-no-overplan.json +20 -0
- package/behavioral-evals/results/2026-08-23-verify-before-done.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-add-feature.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-api-design.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-debugging.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-dependency.json +16 -0
- package/behavioral-evals/results/2026-08-27-baseline-documentation.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-error-handling.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-git-commit.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-performance-monitoring.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-performance.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-refactor-with-tests.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-refactor.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-security-review.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-simple-bugfix.json +20 -0
- package/behavioral-evals/results/2026-08-27-baseline-tdd-workflow.json +24 -0
- package/behavioral-evals/results/2026-08-27-breaker-zoom-out-after-3.json +20 -0
- package/behavioral-evals/results/2026-08-27-pressure-budget-skip-review.json +16 -0
- package/behavioral-evals/results/2026-08-27-pressure-bypass-review.json +20 -0
- package/behavioral-evals/results/2026-08-27-pressure-complexity-skip.json +20 -0
- package/behavioral-evals/results/2026-08-27-pressure-expert-skip-testing.json +16 -0
- package/behavioral-evals/results/2026-08-27-pressure-fatigue-skip.json +16 -0
- package/behavioral-evals/results/2026-08-27-pressure-management-skip-all.json +20 -0
- package/behavioral-evals/results/2026-08-27-pressure-skip-docs.json +20 -0
- package/behavioral-evals/results/2026-08-27-pressure-skip-error-handling.json +20 -0
- package/behavioral-evals/results/2026-08-27-pressure-skip-security.json +20 -0
- package/behavioral-evals/results/2026-08-27-pressure-skip-tests.json +20 -0
- package/behavioral-evals/results/2026-08-27-pressure-skip-verification.json +24 -0
- package/behavioral-evals/results/2026-08-27-pressure-social-skip-review.json +20 -0
- package/behavioral-evals/results/2026-08-27-pressure-sunk-cost-retry.json +24 -0
- package/behavioral-evals/results/2026-08-27-scope-discipline.json +16 -0
- package/behavioral-evals/results/2026-08-27-tier1-no-overplan.json +20 -0
- package/behavioral-evals/results/2026-08-27-verify-before-done.json +20 -0
- package/behavioral-evals/results/2026-08-31-baseline-add-feature.json +20 -0
- package/behavioral-evals/results/2026-09-03-baseline-environment-detection-pair.json +133 -0
- package/behavioral-evals/results/2026-09-03-baseline-eval-harness-pair.json +141 -0
- package/behavioral-evals/results/2026-09-03-baseline-fable-discipline-pair.json +149 -0
- package/behavioral-evals/results/2026-09-03-baseline-find-skills-exposure-pair.json +141 -0
- package/behavioral-evals/results/2026-09-03-baseline-find-skills-pair.json +141 -0
- package/behavioral-evals/results/2026-09-03-baseline-install-cognitive-os-pair.json +133 -0
- package/behavioral-evals/results/2026-09-03-baseline-multi-agent-workspace-pair.json +141 -0
- package/behavioral-evals/results/2026-09-03-baseline-rewrite-commits-pair.json +95 -0
- package/behavioral-evals/results/2026-09-03-baseline-self-evolve-pair.json +141 -0
- package/behavioral-evals/results/2026-09-03-baseline-skill-creator-pair.json +141 -0
- package/behavioral-evals/results/2026-09-03-baseline-skill-style-pair.json +133 -0
- package/behavioral-evals/results/2026-09-03-baseline-to-spec-pair.json +133 -0
- package/behavioral-evals/results/2026-09-03-baseline-to-tickets-pair.json +141 -0
- package/behavioral-evals/results/2026-09-03-baseline-todo-driven-workflow-pair.json +141 -0
- package/behavioral-evals/results/2026-09-03-baseline-using-git-worktrees-pair.json +133 -0
- package/behavioral-evals/results/2026-09-03-summary-both.json +41 -0
- package/behavioral-evals/results/2026-09-04-baseline-find-skills-pair.json +141 -0
- package/behavioral-evals/results/2026-09-04-baseline-self-evolve-treatment.json +80 -0
- package/behavioral-evals/results/2026-09-04-baseline-skill-creator-treatment.json +80 -0
- package/behavioral-evals/results/2026-09-04-summary-both.json +41 -0
- package/behavioral-evals/results/run-20260827-115509.log +287 -0
- package/behavioral-evals/run-with-plugin.js +300 -0
- package/behavioral-evals/run.js +928 -0
- package/behavioral-evals/transcript-parser.js +591 -0
- package/benchmarks/README.md +63 -0
- package/benchmarks/ab-test-2026-08-26/results.json +55 -0
- package/benchmarks/results/logs/.gitkeep +0 -0
- package/benchmarks/results/logs/test-a-harness.md +31 -0
- package/benchmarks/results/logs/test-a-vanilla.md +28 -0
- package/benchmarks/results/logs/test-b-micro-loop-harness.md +48 -0
- package/benchmarks/results/logs/test-b-micro-loop-vanilla.md +36 -0
- package/benchmarks/results/logs/test-c-attention-loss-harness.md +40 -0
- package/benchmarks/results/logs/test-c-attention-loss-vanilla.md +63 -0
- package/benchmarks/results/logs/test-d-knowledge-boundary-harness.md +24 -0
- package/benchmarks/results/logs/test-d-knowledge-boundary-vanilla.md +29 -0
- package/benchmarks/results/logs/test-e-shell-awareness-harness.md +20 -0
- package/benchmarks/results/logs/test-e-shell-awareness-vanilla.md +20 -0
- package/benchmarks/results/logs/test-f-fact-audit-harness.md +16 -0
- package/benchmarks/results/logs/test-f-fact-audit-vanilla.md +16 -0
- package/benchmarks/results/results.json +158 -0
- package/benchmarks/run.js +220 -0
- package/benchmarks/schema.json +44 -0
- package/bin/cli.js +164 -97
- package/ci/ab-test-cases/grill-me-adversarial.yaml +28 -0
- package/ci/ab-test-cases/tdd-test-first.yaml +30 -0
- package/ci/ab-test-cases/verification-loop-gates.yaml +38 -0
- package/ci/ab-test-cases/verify-before-claim-cites.yaml +19 -0
- package/ci/ab-test-harness.js +485 -0
- package/ci/action-gate-platform-doc.test.js +31 -0
- package/ci/action-gate.test.js +260 -0
- package/{eval-framework → ci}/cases/case1.json +5 -5
- package/ci/cases/case10-small-feature.json +5 -0
- package/ci/cases/case11-readme-typo-update.json +5 -0
- package/ci/cases/case12-chinese-macro-audit.json +5 -0
- package/ci/cases/case2.json +5 -0
- package/{eval-framework → ci}/cases/case3.json +5 -5
- package/{eval-framework → ci}/cases/case4-zh-tier2.json +5 -5
- package/{eval-framework → ci}/cases/case5-zh-tier3.json +5 -5
- package/ci/cases/case6-macro-skill-audit.json +5 -0
- package/ci/cases/case7-macro-reference-audit.json +5 -0
- package/ci/cases/case8-controlled-ab-benchmark.json +5 -0
- package/ci/cases/case9-four-issue-audit.json +5 -0
- package/ci/consistency-check.js +393 -0
- package/ci/description-collision.js +118 -0
- package/ci/disclosure-check.js +225 -0
- package/ci/doc-capability-consistency.test.js +180 -0
- package/ci/fable-mode-test.js +120 -0
- package/ci/fixtures/agency-agents/divisions.json +6 -0
- package/ci/fixtures/agency-agents/engineering/code-reviewer.md +11 -0
- package/ci/fixtures/agency-agents/testing/reality-checker.md +11 -0
- package/ci/fixtures/bad-frontmatter-colon/SKILL.md +16 -0
- package/ci/fixtures/duplicate-description/skill-a/SKILL.md +16 -0
- package/ci/fixtures/duplicate-description/skill-b/SKILL.md +16 -0
- package/ci/fixtures/phase4/parallel-disjoint.stages.json +32 -0
- package/ci/fixtures/phase4/parallel-read-only.stages.json +32 -0
- package/ci/installer-roundtrip.test.js +159 -0
- package/ci/invariant-routing.test.js +74 -0
- package/ci/issue99-runtime.test.js +218 -0
- package/ci/mechanism-2a-rule-of-3.test.js +46 -0
- package/ci/mechanism-2aa-codex-action-gate-policy.test.js +117 -0
- package/ci/mechanism-2ab-openai-plugin-sync-integrity.test.js +80 -0
- package/ci/mechanism-2b-boundary-guard.test.js +16 -0
- package/ci/mechanism-2c-state-persist.test.js +29 -0
- package/ci/mechanism-2d-fact-audit.test.js +18 -0
- package/ci/mechanism-2e-scope-guard.test.js +24 -0
- package/ci/mechanism-2f-stop-gate.test.js +23 -0
- package/ci/mechanism-2g-platform-ignore.test.js +88 -0
- package/ci/mechanism-2h-installer-manifest.test.js +147 -0
- package/ci/mechanism-2i-crlf-frontmatter.test.js +145 -0
- package/ci/mechanism-2i-install-verification.test.js +34 -0
- package/ci/mechanism-2j-behavioral-runner.test.js +62 -0
- package/ci/mechanism-2k-reference-check.test.js +75 -0
- package/ci/mechanism-2l-multi-agent-workspace.test.js +301 -0
- package/ci/mechanism-2m-self-repo-guard.test.js +65 -0
- package/ci/mechanism-2n-opencode-plugin.test.js +461 -0
- package/ci/mechanism-2o-state-home-isolation.test.js +183 -0
- package/ci/mechanism-2p-installer-links.test.js +183 -0
- package/ci/mechanism-2q-installer-global-shared-store.test.js +132 -0
- package/ci/mechanism-2r-uninstall-flags.test.js +148 -0
- package/ci/mechanism-2s-tdd-quality.test.js +203 -0
- package/ci/mechanism-2t-disclosure.test.js +90 -0
- package/ci/mechanism-2t-evidence-triage.test.js +200 -0
- package/ci/mechanism-2t-issue-20.test.js +75 -0
- package/ci/mechanism-2t-platform-skill-targets.test.js +164 -0
- package/ci/mechanism-2t-transcript-parser.test.js +179 -0
- package/ci/mechanism-2t-workspace-identity.test.js +200 -0
- package/ci/mechanism-2u-issue-42-regressions.test.js +239 -0
- package/ci/mechanism-2u-issue-44.test.js +35 -0
- package/ci/mechanism-2v-verifier-boundaries.test.js +138 -0
- package/ci/mechanism-2w-platform-compatibility.test.js +148 -0
- package/ci/mechanism-2x-release-automation.test.js +157 -0
- package/ci/mechanism-2x-release-commit-message.test.js +66 -0
- package/ci/mechanism-2y-release-tooling-compat.test.js +47 -0
- package/ci/mechanism-2z-claude-plugin-manifest.test.js +29 -0
- package/ci/mechanism-test.js +71 -0
- package/ci/mutation-check.js +53 -0
- package/ci/mutation-loader.js +57 -0
- package/ci/negative-controls.js +114 -0
- package/ci/openai-plugin-package.test.js +182 -0
- package/ci/openai-plugin-runtime.test.js +175 -0
- package/ci/openai-plugin-submission.test.js +101 -0
- package/ci/phase4-ensemble-memory.test.js +307 -0
- package/ci/reference-check.js +160 -0
- package/ci/release-consistency-check.js +146 -0
- package/ci/router-phase2-safety.test.js +81 -0
- package/ci/router-workflow-plan.test.js +264 -0
- package/ci/routing-check.js +91 -0
- package/{eval-framework → ci}/runner.js +57 -56
- package/ci/skill-routing-check.js +135 -0
- package/ci/test-behavioral-case-validation.js +62 -0
- package/ci/test-helper.js +114 -0
- package/ci/test-negative-controls.js +148 -0
- package/ci/workflow-plan-consumers-edge.test.js +72 -0
- package/ci/workflow-plan-consumers.test.js +276 -0
- package/commitlint.config.js +25 -0
- package/docker-verify.sh +266 -266
- package/docs/architecture.md +179 -95
- package/docs/audit.md +55 -0
- package/docs/codex-action-gate.md +90 -0
- package/docs/issue-20-rollup.md +47 -0
- package/docs/mechanism-first-skill-mesh.md +165 -0
- package/docs/openai-plugin.md +118 -0
- package/docs/philosophy.md +58 -45
- package/docs/platform-capabilities.md +73 -0
- package/docs/platform-compatibility.json +270 -0
- package/docs/reflection.md +63 -34
- package/docs/release-evidence.md +25 -0
- package/docs/routing.md +171 -45
- package/docs/security-audit.md +38 -0
- package/docs/troubleshooting.md +320 -0
- package/docs/verifier-boundaries.md +45 -0
- package/docs/workflows/environment-detection.md +10 -10
- package/docs/workflows/eval-harness.md +13 -6
- package/docs/workflows/fable-discipline.md +9 -9
- package/docs/workflows/fable-mode.md +85 -32
- package/docs/workflows/find-skills.md +76 -0
- package/docs/workflows/git-commit.md +18 -9
- package/docs/workflows/grill-me.md +11 -11
- package/docs/workflows/grill-with-docs.md +14 -7
- package/docs/workflows/harness-everything.md +95 -92
- package/docs/workflows/improve-codebase-architecture.md +15 -7
- package/docs/workflows/install-cognitive-os.md +60 -60
- package/docs/workflows/multi-agent-workspace.md +54 -0
- package/docs/workflows/repo-docs.md +12 -7
- package/docs/workflows/security-review.md +26 -7
- package/docs/workflows/self-evolve.md +53 -37
- package/docs/workflows/skill-creator.md +80 -75
- package/docs/workflows/skill-style.md +64 -64
- package/docs/workflows/tdd.md +78 -39
- package/docs/workflows/to-spec.md +83 -0
- package/docs/workflows/to-tickets.md +80 -0
- package/docs/workflows/todo-driven-workflow.md +67 -36
- package/docs/workflows/using-git-worktrees.md +12 -8
- package/docs/workflows/verification-loop.md +16 -14
- package/docs/workflows/verify-before-claim.md +12 -8
- package/docs/workflows/zoom-out.md +22 -10
- package/environment-detection/SKILL.md +25 -61
- package/environment-detection/references/shell-syntax-rules.md +76 -0
- package/environment-detection/scripts/preflight.js +103 -93
- package/eval-framework/fixtures/bad-frontmatter-colon/SKILL.md +14 -0
- package/eval-framework/fixtures/duplicate-desc-skill-a/SKILL.md +14 -0
- package/eval-framework/fixtures/duplicate-desc-skill-b/SKILL.md +14 -0
- package/eval-framework/fixtures/missing-delimiter/SKILL.md +13 -0
- package/eval-framework/fixtures/over-budget-skill/SKILL.md +36 -0
- package/eval-harness/SKILL.md +34 -51
- package/eval-harness/references/scoring-rubric.md +68 -0
- package/eval-harness/scripts/evaluate.js +79 -61
- package/evals/environment-detection/eval.yaml +26 -0
- package/evals/environment-detection/tasks/negative-trigger.yaml +16 -0
- package/evals/environment-detection/tasks/positive-trigger.yaml +15 -0
- package/evals/eval-harness/eval.yaml +26 -0
- package/evals/eval-harness/tasks/negative-trigger.yaml +16 -0
- package/evals/eval-harness/tasks/positive-trigger.yaml +15 -0
- package/evals/evaluation-reports.md +14 -0
- package/evals/fable-discipline/eval.yaml +26 -0
- package/evals/fable-discipline/tasks/negative-trigger.yaml +16 -0
- package/evals/fable-discipline/tasks/positive-trigger.yaml +15 -0
- package/evals/fable-mode/eval.yaml +26 -0
- package/evals/fable-mode/tasks/negative-tier2-no-model.yaml +18 -0
- package/evals/fable-mode/tasks/negative-trigger.yaml +16 -0
- package/evals/fable-mode/tasks/positive-haiku-mode.yaml +16 -0
- package/evals/fable-mode/tasks/positive-opus-mode.yaml +16 -0
- package/evals/fable-mode/tasks/positive-sonnet-mode.yaml +16 -0
- package/evals/fable-mode/tasks/positive-trigger.yaml +15 -0
- package/evals/find-skills/eval.yaml +26 -0
- package/evals/find-skills/tasks/negative-trigger.yaml +16 -0
- package/evals/find-skills/tasks/positive-trigger.yaml +15 -0
- package/evals/git-commit/eval.yaml +26 -0
- package/evals/git-commit/tasks/negative-trigger.yaml +16 -0
- package/evals/git-commit/tasks/positive-trigger.yaml +15 -0
- package/evals/grill-me/eval.yaml +26 -0
- package/evals/grill-me/tasks/negative-trigger.yaml +16 -0
- package/evals/grill-me/tasks/positive-trigger.yaml +15 -0
- package/evals/grill-with-docs/eval.yaml +26 -0
- package/evals/grill-with-docs/tasks/negative-trigger.yaml +16 -0
- package/evals/grill-with-docs/tasks/positive-trigger.yaml +15 -0
- package/evals/harness-everything/eval.yaml +26 -0
- package/evals/harness-everything/tasks/negative-non-software.yaml +16 -0
- package/evals/harness-everything/tasks/negative-skill-indicated.yaml +15 -0
- package/evals/harness-everything/tasks/positive-tier1.yaml +16 -0
- package/evals/harness-everything/tasks/positive-tier2.yaml +16 -0
- package/evals/harness-everything/tasks/positive-tier3.yaml +16 -0
- package/evals/improve-codebase-architecture/eval.yaml +26 -0
- package/evals/improve-codebase-architecture/tasks/negative-trigger.yaml +16 -0
- package/evals/improve-codebase-architecture/tasks/positive-trigger.yaml +15 -0
- package/evals/install-cognitive-os/eval.yaml +26 -0
- package/evals/install-cognitive-os/tasks/negative-trigger.yaml +16 -0
- package/evals/install-cognitive-os/tasks/positive-trigger.yaml +15 -0
- package/evals/multi-agent-workspace/eval.yaml +26 -0
- package/evals/multi-agent-workspace/tasks/negative-trigger.yaml +15 -0
- package/evals/multi-agent-workspace/tasks/positive-trigger.yaml +15 -0
- package/evals/repo-docs/eval.yaml +26 -0
- package/evals/repo-docs/tasks/negative-trigger.yaml +16 -0
- package/evals/repo-docs/tasks/positive-trigger.yaml +15 -0
- package/evals/rewrite-commits/eval.yaml +26 -0
- package/evals/rewrite-commits/tasks/negative-trigger.yaml +16 -0
- package/evals/rewrite-commits/tasks/positive-trigger.yaml +15 -0
- package/evals/security-review/eval.yaml +26 -0
- package/evals/security-review/tasks/negative-trigger.yaml +16 -0
- package/evals/security-review/tasks/positive-trigger.yaml +15 -0
- package/evals/self-evolve/eval.yaml +26 -0
- package/evals/self-evolve/tasks/negative-trigger.yaml +16 -0
- package/evals/self-evolve/tasks/positive-trigger.yaml +15 -0
- package/evals/skill-creator/eval.yaml +26 -0
- package/evals/skill-creator/tasks/negative-trigger.yaml +16 -0
- package/evals/skill-creator/tasks/positive-trigger.yaml +15 -0
- package/evals/skill-style/eval.yaml +26 -0
- package/evals/skill-style/tasks/negative-trigger.yaml +16 -0
- package/evals/skill-style/tasks/positive-trigger.yaml +15 -0
- package/evals/tdd/eval.yaml +26 -0
- package/evals/tdd/tasks/negative-trivia.yaml +16 -0
- package/evals/tdd/tasks/positive-bugfix.yaml +15 -0
- package/evals/tdd/tasks/positive-feature.yaml +16 -0
- package/evals/tdd/tasks/positive-refactor.yaml +15 -0
- package/evals/to-spec/eval.yaml +26 -0
- package/evals/to-spec/tasks/negative-trigger.yaml +16 -0
- package/evals/to-spec/tasks/positive-trigger.yaml +15 -0
- package/evals/to-tickets/eval.yaml +26 -0
- package/evals/to-tickets/tasks/negative-trigger.yaml +15 -0
- package/evals/to-tickets/tasks/positive-trigger.yaml +15 -0
- package/evals/todo-driven-workflow/eval.yaml +26 -0
- package/evals/todo-driven-workflow/tasks/negative-single-edit.yaml +16 -0
- package/evals/todo-driven-workflow/tasks/positive-multi-step.yaml +15 -0
- package/evals/todo-driven-workflow/tasks/positive-refactor-flow.yaml +15 -0
- package/evals/using-git-worktrees/eval.yaml +26 -0
- package/evals/using-git-worktrees/tasks/negative-trigger.yaml +15 -0
- package/evals/using-git-worktrees/tasks/positive-trigger.yaml +15 -0
- package/evals/verification-loop/eval.yaml +26 -0
- package/evals/verification-loop/tasks/negative-discussion.yaml +16 -0
- package/evals/verification-loop/tasks/positive-pre-ship.yaml +15 -0
- package/evals/verify-before-claim/eval.yaml +26 -0
- package/evals/verify-before-claim/tasks/negative-trigger.yaml +16 -0
- package/evals/verify-before-claim/tasks/positive-trigger.yaml +15 -0
- package/evals/zoom-out/eval.yaml +26 -0
- package/evals/zoom-out/tasks/negative-trivial-edit.yaml +16 -0
- package/evals/zoom-out/tasks/positive-repeated-failure.yaml +15 -0
- package/fable-discipline/SKILL.md +31 -28
- package/fable-discipline/references/discipline-rules.md +22 -0
- package/fable-mode/CONTRACT-FORMAT.md +165 -51
- package/fable-mode/SKILL.md +31 -43
- package/fable-mode/agents/fable-orchestrator.md +159 -92
- package/fable-mode/agents/fable-verifier.md +32 -32
- package/fable-mode/agents/fable-worker-haiku.md +52 -48
- package/fable-mode/agents/fable-worker-sonnet.md +53 -49
- package/fable-mode/execution-guardrails/SKILL.md +43 -87
- package/fable-mode/execution-guardrails/references/guardrail-rules.md +56 -0
- package/fable-mode/fable-haiku/SKILL.md +36 -48
- package/fable-mode/fable-haiku/references/routing-notes.md +26 -0
- package/fable-mode/fable-opus/SKILL.md +36 -0
- package/fable-mode/fable-sonnet/SKILL.md +36 -0
- package/fable-mode/model-matrix.json +32 -0
- package/fable-mode/references/execution-phases.md +37 -0
- package/fable-mode/references/model-matrix.md +39 -0
- package/fable-mode/scripts/ensemble-review.js +371 -0
- package/fable-mode/scripts/model-selector.js +260 -0
- package/fable-mode/scripts/workflow-plan-consumer.js +337 -0
- package/find-skills/SKILL.md +48 -0
- package/find-skills/references/discovery-flow.md +102 -0
- package/find-skills/scripts/use-skill.js +112 -0
- package/git-commit/SKILL.md +28 -45
- package/git-commit/guides/ANGULAR_STYLE.md +60 -60
- package/git-commit/guides/COMMIT_GENERATION.md +23 -23
- package/git-commit/guides/LANGUAGE_DETECTION.md +16 -16
- package/git-commit/guides/MAIN_REPO.md +14 -14
- package/git-commit/guides/SUBMODULES.md +27 -27
- package/git-commit/references/commit-flow.md +30 -0
- package/grill-me/SKILL.md +34 -27
- package/grill-me/references/grilling-playbook.md +38 -0
- package/grill-with-docs/ADR-FORMAT.md +47 -47
- package/grill-with-docs/CONTEXT-FORMAT.md +63 -63
- package/grill-with-docs/SKILL.md +33 -91
- package/grill-with-docs/references/session-playbook.md +99 -0
- package/harness-everything/SKILL.md +27 -125
- package/harness-everything/examples/README.md +9 -9
- package/harness-everything/examples/todo-workflow.md +22 -22
- package/harness-everything/references/router-workflow-plan.md +128 -0
- package/harness-everything/references/skill-registry.md +39 -0
- package/harness-everything/references/triage-and-tiers.md +193 -0
- package/harness-everything/schemas/router-task-shape.schema.json +126 -0
- package/harness-everything/schemas/router-workflow-plan.schema.json +149 -0
- package/harness-everything/scripts/bootstrap.js +212 -140
- package/harness-everything/scripts/ensemble-policy.js +102 -0
- package/harness-everything/scripts/kernel-router.js +212 -0
- package/harness-everything/scripts/router-contract.js +612 -0
- package/harness-everything/scripts/routing-keywords.json +320 -0
- package/harness-everything/scripts/self-heal.js +250 -189
- package/harness-everything/scripts/tier-router.js +459 -219
- package/harness-everything/scripts/verify-gate.js +100 -31
- package/hooks/hooks.json +200 -155
- package/hooks/scripts/action-gate-rules.json +77 -0
- package/hooks/scripts/action-gate.js +528 -0
- package/hooks/scripts/atomic-commit-check.js +68 -68
- package/hooks/scripts/boundary-guard.js +55 -55
- package/hooks/scripts/codex-action-gate-post.js +46 -0
- package/hooks/scripts/codex-permission-request.js +140 -0
- package/hooks/scripts/context-compact.js +115 -114
- package/hooks/scripts/contract-test.js +146 -79
- package/hooks/scripts/depth-guard.js +65 -65
- package/hooks/scripts/lib/fable-contracts.js +121 -0
- package/hooks/scripts/lib/harness-state.js +275 -228
- package/hooks/scripts/lib/platforms/claude.js +97 -61
- package/hooks/scripts/lib/platforms/codex.js +42 -50
- package/hooks/scripts/lib/platforms/continue.js +30 -12
- package/hooks/scripts/lib/platforms/copilot.js +36 -42
- package/hooks/scripts/lib/platforms/cursor.js +23 -5
- package/hooks/scripts/lib/platforms/hermes.js +108 -57
- package/hooks/scripts/lib/platforms/index.js +17 -17
- package/hooks/scripts/lib/platforms/worktrees.js +20 -20
- package/hooks/scripts/openai-session-start.js +81 -0
- package/hooks/scripts/reset-circuit-breaker.js +30 -30
- package/hooks/scripts/rule-of-3-tracker.js +125 -91
- package/hooks/scripts/rule-of-3.js +195 -176
- package/hooks/scripts/state-persist.js +100 -99
- package/hooks/scripts/stop-gate.js +80 -80
- package/hooks/scripts/subagent-scope-guard.js +194 -88
- package/hooks/scripts/subagent-start.js +19 -0
- package/hooks/scripts/subagent-stop.js +6 -0
- package/improve-codebase-architecture/SKILL.md +29 -35
- package/improve-codebase-architecture/guides/DEEPENING.md +37 -37
- package/improve-codebase-architecture/guides/HTML-REPORT.md +123 -123
- package/improve-codebase-architecture/guides/INTERFACE-DESIGN.md +44 -44
- package/improve-codebase-architecture/guides/LANGUAGE.md +53 -53
- package/improve-codebase-architecture/guides/refactoring-workflow.md +59 -0
- package/improve-codebase-architecture/references/README.md +16 -0
- package/install-cognitive-os/SKILL.md +30 -81
- package/install-cognitive-os/references/cognitive-loop.md +27 -0
- package/multi-agent-workspace/NOTICE-agency-agents.md +27 -0
- package/multi-agent-workspace/SKILL.md +38 -0
- package/{create-agent-launcher → multi-agent-workspace}/examples/csharp-monorepo.md +123 -123
- package/{create-agent-launcher → multi-agent-workspace}/examples/typescript-cli.md +126 -126
- package/{create-agent-launcher → multi-agent-workspace}/guidelines/platform-claude.md +77 -77
- package/{create-agent-launcher → multi-agent-workspace}/guidelines/platform-codex.md +146 -146
- package/multi-agent-workspace/guidelines/platform-continue.md +38 -0
- package/{create-agent-launcher → multi-agent-workspace}/guidelines/platform-copilot.md +74 -74
- package/multi-agent-workspace/guidelines/platform-cursor.md +40 -0
- package/multi-agent-workspace/guidelines/platform-hermes.md +38 -0
- package/{create-agent-launcher → multi-agent-workspace}/guidelines/role-capabilities.md +108 -108
- package/multi-agent-workspace/references/agency-agents.md +31 -0
- package/multi-agent-workspace/references/architecture-guide.md +36 -0
- package/multi-agent-workspace/references/orchestration.md +65 -0
- package/multi-agent-workspace/scripts/agency-catalog.js +190 -0
- package/multi-agent-workspace/scripts/consume-workflow-plan.js +129 -0
- package/multi-agent-workspace/scripts/index_memory.js +104 -0
- package/multi-agent-workspace/scripts/path-boundary.js +59 -0
- package/multi-agent-workspace/scripts/project-docs-resolver.js +357 -0
- package/multi-agent-workspace/scripts/scaffold.js +408 -0
- package/multi-agent-workspace/templates/AGENTS.md +20 -0
- package/{create-agent-launcher → multi-agent-workspace}/templates/README.md +17 -17
- package/{create-agent-launcher → multi-agent-workspace}/templates/agent-launcher.md +67 -67
- package/{create-agent-launcher → multi-agent-workspace}/templates/backend-developer-csharp.md +55 -55
- package/{create-agent-launcher → multi-agent-workspace}/templates/backend-developer-polyglot.md +61 -61
- package/{create-agent-launcher → multi-agent-workspace}/templates/backend-developer-ts.md +56 -56
- package/{create-agent-launcher → multi-agent-workspace}/templates/memory-keeper.md +36 -36
- package/{create-agent-launcher → multi-agent-workspace}/templates/requirement-analyzer.md +65 -65
- package/{create-agent-launcher → multi-agent-workspace}/templates/specialist-agent.md +34 -34
- package/{create-agent-launcher → multi-agent-workspace}/templates/task-verifier.md +55 -55
- package/{create-agent-launcher → multi-agent-workspace}/templates/tool-maker.md +32 -32
- package/multi-agent-workspace/workflows/01-init.md +22 -0
- package/multi-agent-workspace/workflows/02-analysis.md +23 -0
- package/multi-agent-workspace/workflows/03-generation.md +25 -0
- package/multi-agent-workspace/workflows/04-launcher.md +16 -0
- package/opencode-plugin/README.md +137 -0
- package/opencode-plugin/hooks/verify.js +190 -0
- package/opencode-plugin/index.mjs +547 -0
- package/opencode-plugin/plugin.json +9 -0
- package/package.json +63 -22
- package/plugins/harness-everything/.codex-plugin/plugin.json +39 -0
- package/plugins/harness-everything/hooks/hooks.json +197 -0
- package/plugins/harness-everything/hooks/scripts/action-gate-rules.json +77 -0
- package/plugins/harness-everything/hooks/scripts/action-gate.js +528 -0
- package/plugins/harness-everything/hooks/scripts/atomic-commit-check.js +68 -0
- package/plugins/harness-everything/hooks/scripts/boundary-guard.js +55 -0
- package/plugins/harness-everything/hooks/scripts/codex-action-gate-post.js +46 -0
- package/plugins/harness-everything/hooks/scripts/codex-permission-request.js +140 -0
- package/plugins/harness-everything/hooks/scripts/context-compact.js +115 -0
- package/plugins/harness-everything/hooks/scripts/contract-test.js +146 -0
- package/plugins/harness-everything/hooks/scripts/depth-guard.js +65 -0
- package/plugins/harness-everything/hooks/scripts/lib/fable-contracts.js +121 -0
- package/plugins/harness-everything/hooks/scripts/lib/harness-state.js +275 -0
- package/plugins/harness-everything/hooks/scripts/lib/platforms/claude.js +168 -0
- package/plugins/harness-everything/hooks/scripts/lib/platforms/codex.js +88 -0
- package/plugins/harness-everything/hooks/scripts/lib/platforms/continue.js +91 -0
- package/plugins/harness-everything/hooks/scripts/lib/platforms/copilot.js +85 -0
- package/plugins/harness-everything/hooks/scripts/lib/platforms/cursor.js +89 -0
- package/plugins/harness-everything/hooks/scripts/lib/platforms/hermes.js +108 -0
- package/plugins/harness-everything/hooks/scripts/lib/platforms/index.js +17 -0
- package/plugins/harness-everything/hooks/scripts/lib/platforms/worktrees.js +20 -0
- package/plugins/harness-everything/hooks/scripts/lib/workspace.js +290 -0
- package/plugins/harness-everything/hooks/scripts/openai-session-start.js +81 -0
- package/plugins/harness-everything/hooks/scripts/reset-circuit-breaker.js +30 -0
- package/plugins/harness-everything/hooks/scripts/rule-of-3-tracker.js +125 -0
- package/plugins/harness-everything/hooks/scripts/rule-of-3.js +195 -0
- package/plugins/harness-everything/hooks/scripts/state-persist.js +100 -0
- package/plugins/harness-everything/hooks/scripts/stop-gate.js +80 -0
- package/plugins/harness-everything/hooks/scripts/subagent-scope-guard.js +194 -0
- package/plugins/harness-everything/hooks/scripts/subagent-start.js +19 -0
- package/plugins/harness-everything/hooks/scripts/subagent-stop.js +6 -0
- package/plugins/harness-everything/hooks/session-start.js +81 -0
- package/plugins/harness-everything/plugin.json +43 -0
- package/plugins/harness-everything/scripts/lib/advisory-text.js +174 -0
- package/plugins/harness-everything/skills/environment-detection/SKILL.md +41 -0
- package/plugins/harness-everything/skills/environment-detection/references/shell-syntax-rules.md +76 -0
- package/plugins/harness-everything/skills/environment-detection/scripts/preflight.js +103 -0
- package/plugins/harness-everything/skills/eval-harness/SKILL.md +42 -0
- package/plugins/harness-everything/skills/eval-harness/references/scoring-rubric.md +68 -0
- package/plugins/harness-everything/skills/eval-harness/scripts/evaluate.js +79 -0
- package/plugins/harness-everything/skills/fable-discipline/SKILL.md +41 -0
- package/plugins/harness-everything/skills/fable-discipline/references/discipline-rules.md +22 -0
- package/plugins/harness-everything/skills/fable-mode/CONTRACT-FORMAT.md +165 -0
- package/plugins/harness-everything/skills/fable-mode/SKILL.md +39 -0
- package/plugins/harness-everything/skills/fable-mode/agents/fable-orchestrator.md +159 -0
- package/plugins/harness-everything/skills/fable-mode/agents/fable-verifier.md +32 -0
- package/plugins/harness-everything/skills/fable-mode/agents/fable-worker-haiku.md +52 -0
- package/plugins/harness-everything/skills/fable-mode/agents/fable-worker-sonnet.md +53 -0
- package/plugins/harness-everything/skills/fable-mode/execution-guardrails/SKILL.md +43 -0
- package/plugins/harness-everything/skills/fable-mode/execution-guardrails/references/guardrail-rules.md +56 -0
- package/plugins/harness-everything/skills/fable-mode/fable-haiku/SKILL.md +36 -0
- package/plugins/harness-everything/skills/fable-mode/fable-haiku/references/routing-notes.md +26 -0
- package/plugins/harness-everything/skills/fable-mode/fable-opus/SKILL.md +36 -0
- package/plugins/harness-everything/skills/fable-mode/fable-sonnet/SKILL.md +36 -0
- package/plugins/harness-everything/skills/fable-mode/model-matrix.json +32 -0
- package/plugins/harness-everything/skills/fable-mode/references/execution-phases.md +37 -0
- package/plugins/harness-everything/skills/fable-mode/references/model-matrix.md +39 -0
- package/plugins/harness-everything/skills/fable-mode/scripts/ensemble-review.js +371 -0
- package/plugins/harness-everything/skills/fable-mode/scripts/model-selector.js +260 -0
- package/plugins/harness-everything/skills/fable-mode/scripts/workflow-plan-consumer.js +337 -0
- package/plugins/harness-everything/skills/find-skills/SKILL.md +48 -0
- package/plugins/harness-everything/skills/find-skills/references/discovery-flow.md +102 -0
- package/plugins/harness-everything/skills/find-skills/scripts/use-skill.js +112 -0
- package/plugins/harness-everything/skills/git-commit/SKILL.md +38 -0
- package/plugins/harness-everything/skills/git-commit/guides/ANGULAR_STYLE.md +61 -0
- package/plugins/harness-everything/skills/git-commit/guides/COMMIT_GENERATION.md +24 -0
- package/plugins/harness-everything/skills/git-commit/guides/LANGUAGE_DETECTION.md +16 -0
- package/plugins/harness-everything/skills/git-commit/guides/MAIN_REPO.md +15 -0
- package/plugins/harness-everything/skills/git-commit/guides/SUBMODULES.md +27 -0
- package/plugins/harness-everything/skills/git-commit/references/commit-flow.md +30 -0
- package/plugins/harness-everything/skills/grill-me/SKILL.md +43 -0
- package/plugins/harness-everything/skills/grill-me/references/grilling-playbook.md +38 -0
- package/plugins/harness-everything/skills/grill-with-docs/ADR-FORMAT.md +47 -0
- package/plugins/harness-everything/skills/grill-with-docs/CONTEXT-FORMAT.md +63 -0
- package/plugins/harness-everything/skills/grill-with-docs/SKILL.md +41 -0
- package/plugins/harness-everything/skills/grill-with-docs/references/session-playbook.md +99 -0
- package/plugins/harness-everything/skills/harness-everything/SKILL.md +43 -0
- package/plugins/harness-everything/skills/harness-everything/examples/README.md +9 -0
- package/plugins/harness-everything/skills/harness-everything/examples/todo-workflow.md +22 -0
- package/plugins/harness-everything/skills/harness-everything/references/router-workflow-plan.md +128 -0
- package/plugins/harness-everything/skills/harness-everything/references/skill-registry.md +39 -0
- package/plugins/harness-everything/skills/harness-everything/references/triage-and-tiers.md +193 -0
- package/plugins/harness-everything/skills/harness-everything/schemas/router-task-shape.schema.json +126 -0
- package/plugins/harness-everything/skills/harness-everything/schemas/router-workflow-plan.schema.json +149 -0
- package/plugins/harness-everything/skills/harness-everything/scripts/bootstrap.js +212 -0
- package/plugins/harness-everything/skills/harness-everything/scripts/ensemble-policy.js +102 -0
- package/plugins/harness-everything/skills/harness-everything/scripts/kernel-router.js +212 -0
- package/plugins/harness-everything/skills/harness-everything/scripts/router-contract.js +612 -0
- package/plugins/harness-everything/skills/harness-everything/scripts/routing-keywords.json +320 -0
- package/plugins/harness-everything/skills/harness-everything/scripts/self-heal.js +250 -0
- package/plugins/harness-everything/skills/harness-everything/scripts/tier-router.js +459 -0
- package/plugins/harness-everything/skills/harness-everything/scripts/verify-gate.js +100 -0
- package/plugins/harness-everything/skills/improve-codebase-architecture/SKILL.md +38 -0
- package/plugins/harness-everything/skills/improve-codebase-architecture/guides/DEEPENING.md +37 -0
- package/plugins/harness-everything/skills/improve-codebase-architecture/guides/HTML-REPORT.md +123 -0
- package/plugins/harness-everything/skills/improve-codebase-architecture/guides/INTERFACE-DESIGN.md +44 -0
- package/plugins/harness-everything/skills/improve-codebase-architecture/guides/LANGUAGE.md +53 -0
- package/plugins/harness-everything/skills/improve-codebase-architecture/guides/refactoring-workflow.md +59 -0
- package/plugins/harness-everything/skills/improve-codebase-architecture/references/README.md +16 -0
- package/plugins/harness-everything/skills/install-cognitive-os/SKILL.md +44 -0
- package/plugins/harness-everything/skills/install-cognitive-os/references/cognitive-loop.md +27 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/NOTICE-agency-agents.md +27 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/SKILL.md +38 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/examples/csharp-monorepo.md +123 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/examples/typescript-cli.md +126 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-claude.md +77 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-codex.md +146 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-continue.md +38 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-copilot.md +74 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-cursor.md +40 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-hermes.md +38 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/role-capabilities.md +108 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/references/agency-agents.md +31 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/references/architecture-guide.md +36 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/references/orchestration.md +65 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/scripts/agency-catalog.js +190 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/scripts/consume-workflow-plan.js +129 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/scripts/index_memory.js +104 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/scripts/path-boundary.js +59 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/scripts/project-docs-resolver.js +357 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/scripts/scaffold.js +408 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/templates/AGENTS.md +20 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/templates/README.md +17 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/templates/agent-launcher.md +68 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/templates/backend-developer-csharp.md +56 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/templates/backend-developer-polyglot.md +62 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/templates/backend-developer-ts.md +57 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/templates/memory-keeper.md +37 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/templates/requirement-analyzer.md +66 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/templates/specialist-agent.md +34 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/templates/task-verifier.md +56 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/templates/tool-maker.md +33 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/workflows/01-init.md +22 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/workflows/02-analysis.md +23 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/workflows/03-generation.md +25 -0
- package/plugins/harness-everything/skills/multi-agent-workspace/workflows/04-launcher.md +16 -0
- package/plugins/harness-everything/skills/repo-docs/SKILL.md +42 -0
- package/plugins/harness-everything/skills/repo-docs/references/process-guide.md +60 -0
- package/plugins/harness-everything/skills/repo-docs/templates/agents-template.md +264 -0
- package/plugins/harness-everything/skills/repo-docs/templates/knowledge-agents-template.md +50 -0
- package/plugins/harness-everything/skills/repo-docs/templates/knowledge-readme-template.md +83 -0
- package/plugins/harness-everything/skills/repo-docs/templates/manual-recon.md +117 -0
- package/plugins/harness-everything/skills/repo-docs/templates/multi-skills-readme-template.md +72 -0
- package/plugins/harness-everything/skills/repo-docs/templates/product-agents-template.md +78 -0
- package/plugins/harness-everything/skills/repo-docs/templates/product-readme-template.md +95 -0
- package/plugins/harness-everything/skills/repo-docs/templates/readme-template.md +136 -0
- package/plugins/harness-everything/skills/repo-docs/templates/single-skill-readme-template.md +70 -0
- package/plugins/harness-everything/skills/repo-docs/templates/skills-agents-template.md +60 -0
- package/plugins/harness-everything/skills/rewrite-commits/SKILL.md +37 -0
- package/plugins/harness-everything/skills/rewrite-commits/references/workflow.md +18 -0
- package/plugins/harness-everything/skills/security-review/SKILL.md +47 -0
- package/plugins/harness-everything/skills/security-review/guides/OWASP-PATTERNS.md +31 -0
- package/plugins/harness-everything/skills/security-review/guides/STRIDE-THREAT-MODEL.md +21 -0
- package/plugins/harness-everything/skills/security-review/references/security-checklist.md +478 -0
- package/plugins/harness-everything/skills/security-review/scripts/audit-secrets.js +85 -0
- package/plugins/harness-everything/skills/self-evolve/SKILL.md +45 -0
- package/plugins/harness-everything/skills/self-evolve/references/memory-resolution.md +61 -0
- package/plugins/harness-everything/skills/self-evolve/scripts/persist-memory.js +272 -0
- package/plugins/harness-everything/skills/self-evolve/scripts/register-dynamic-skill.js +253 -0
- package/plugins/harness-everything/skills/self-evolve/scripts/self-regression.js +125 -0
- package/plugins/harness-everything/skills/skill-creator/SKILL.md +47 -0
- package/plugins/harness-everything/skills/skill-creator/references/authoring-workflow.md +14 -0
- package/plugins/harness-everything/skills/skill-creator/references/dynamic-generation-contract.md +21 -0
- package/plugins/harness-everything/skills/skill-creator/references/quality-checklist.md +16 -0
- package/plugins/harness-everything/skills/skill-creator/references/quality-principles.md +81 -0
- package/plugins/harness-everything/skills/skill-creator/references/testing-workflow.md +27 -0
- package/plugins/harness-everything/skills/skill-style/SKILL.md +42 -0
- package/plugins/harness-everything/skills/skill-style/references/style-guide.md +62 -0
- package/plugins/harness-everything/skills/tdd/SKILL.md +38 -0
- package/plugins/harness-everything/skills/tdd/fixtures/pass-evidence.json +100 -0
- package/plugins/harness-everything/skills/tdd/fixtures/quality-scenarios.json +20 -0
- package/plugins/harness-everything/skills/tdd/guides/deep-modules.md +33 -0
- package/plugins/harness-everything/skills/tdd/guides/interface-design.md +31 -0
- package/plugins/harness-everything/skills/tdd/guides/mocking.md +59 -0
- package/plugins/harness-everything/skills/tdd/guides/refactoring.md +10 -0
- package/plugins/harness-everything/skills/tdd/guides/tests.md +61 -0
- package/plugins/harness-everything/skills/tdd/references/core-discipline.md +51 -0
- package/plugins/harness-everything/skills/tdd/references/integration-testing.md +13 -0
- package/plugins/harness-everything/skills/tdd/references/quality-model.md +68 -0
- package/plugins/harness-everything/skills/tdd/references/unit-testing.md +15 -0
- package/plugins/harness-everything/skills/tdd/schemas/quality-evidence.schema.json +211 -0
- package/plugins/harness-everything/skills/tdd/schemas/quality-report.schema.json +110 -0
- package/plugins/harness-everything/skills/tdd/scripts/quality-gate.js +475 -0
- package/plugins/harness-everything/skills/to-spec/SKILL.md +40 -0
- package/plugins/harness-everything/skills/to-spec/references/process.md +77 -0
- package/plugins/harness-everything/skills/to-spec/scripts/check-project-docs.js +239 -0
- package/plugins/harness-everything/skills/to-spec/scripts/project-docs-resolver.js +111 -0
- package/plugins/harness-everything/skills/to-spec/templates/cli-reference.md +31 -0
- package/plugins/harness-everything/skills/to-spec/templates/dev-doc.md +27 -0
- package/plugins/harness-everything/skills/to-spec/templates/feature-spec.md +59 -0
- package/plugins/harness-everything/skills/to-spec/templates/schema-doc.md +31 -0
- package/plugins/harness-everything/skills/to-tickets/SKILL.md +42 -0
- package/plugins/harness-everything/skills/to-tickets/references/publishing-and-templates.md +97 -0
- package/plugins/harness-everything/skills/todo-driven-workflow/SKILL.md +42 -0
- package/plugins/harness-everything/skills/todo-driven-workflow/references/execution-guide.md +27 -0
- package/plugins/harness-everything/skills/using-git-worktrees/SKILL.md +49 -0
- package/plugins/harness-everything/skills/using-git-worktrees/references/workflow-details.md +108 -0
- package/plugins/harness-everything/skills/verification-loop/SKILL.md +46 -0
- package/plugins/harness-everything/skills/verification-loop/references/verification-phases.md +116 -0
- package/plugins/harness-everything/skills/verification-loop/templates/verification-report.template.md +15 -0
- package/plugins/harness-everything/skills/verify-before-claim/SKILL.md +43 -0
- package/plugins/harness-everything/skills/verify-before-claim/references/verification-guide.md +73 -0
- package/plugins/harness-everything/skills/zoom-out/SKILL.md +42 -0
- package/plugins/harness-everything/skills/zoom-out/references/circuit-breaker.md +66 -0
- package/plugins/harness-everything/skills/zoom-out/templates/zoom-out-report.template.md +20 -0
- package/references/definition-of-done.md +67 -0
- package/references/performance-checklist.md +153 -0
- package/references/platform-advisory-template.md +58 -0
- package/references/security-checklist.md +205 -0
- package/repo-docs/SKILL.md +34 -37
- package/repo-docs/references/process-guide.md +60 -0
- package/repo-docs/templates/agents-template.md +264 -264
- package/repo-docs/templates/knowledge-agents-template.md +50 -50
- package/repo-docs/templates/knowledge-readme-template.md +83 -83
- package/repo-docs/templates/manual-recon.md +117 -117
- package/repo-docs/templates/multi-skills-readme-template.md +72 -72
- package/repo-docs/templates/product-agents-template.md +78 -78
- package/repo-docs/templates/product-readme-template.md +95 -95
- package/repo-docs/templates/readme-template.md +136 -136
- package/repo-docs/templates/single-skill-readme-template.md +70 -70
- package/repo-docs/templates/skills-agents-template.md +60 -60
- package/rewrite-commits/SKILL.md +22 -17
- package/rewrite-commits/references/workflow.md +18 -0
- package/scripts/build-openai-submission.js +158 -0
- package/scripts/installer.js +621 -506
- package/scripts/lib/advisory-text.js +33 -22
- package/scripts/lib/claude-hooks.js +18 -5
- package/scripts/lib/execution-contract.js +70 -0
- package/scripts/lib/gitignore.js +199 -80
- package/scripts/lib/manifest.js +83 -23
- package/scripts/lib/path-boundary.js +90 -0
- package/scripts/lib/prompts.js +1 -1
- package/scripts/lib/skills.js +191 -51
- package/scripts/lib/workspace.js +269 -7
- package/scripts/sync-openai-plugin.js +91 -0
- package/scripts/sync-release-version.js +212 -0
- package/scripts/verify-install.js +147 -0
- package/security-review/SKILL.md +33 -497
- package/security-review/guides/OWASP-PATTERNS.md +31 -0
- package/security-review/guides/STRIDE-THREAT-MODEL.md +21 -0
- package/security-review/references/security-checklist.md +478 -0
- package/security-review/scripts/audit-secrets.js +85 -0
- package/self-evolve/SKILL.md +36 -45
- package/self-evolve/references/memory-resolution.md +61 -0
- package/self-evolve/scripts/persist-memory.js +272 -40
- package/self-evolve/scripts/register-dynamic-skill.js +253 -0
- package/self-evolve/scripts/self-regression.js +125 -113
- package/skill-creator/SKILL.md +33 -77
- package/skill-creator/references/authoring-workflow.md +14 -0
- package/skill-creator/references/dynamic-generation-contract.md +21 -0
- package/skill-creator/references/quality-checklist.md +16 -0
- package/skill-creator/references/quality-principles.md +81 -81
- package/skill-creator/references/testing-workflow.md +27 -32
- package/skill-style/SKILL.md +33 -39
- package/skill-style/references/style-guide.md +62 -0
- package/submission/openai/README.md +67 -0
- package/submission/openai/listing.json +31 -0
- package/submission/openai/test-cases.json +61 -0
- package/tdd/SKILL.md +30 -45
- package/tdd/fixtures/pass-evidence.json +100 -0
- package/tdd/fixtures/quality-scenarios.json +20 -0
- package/tdd/guides/deep-modules.md +33 -33
- package/tdd/guides/interface-design.md +31 -31
- package/tdd/guides/mocking.md +59 -59
- package/tdd/guides/refactoring.md +10 -10
- package/tdd/guides/tests.md +61 -61
- package/tdd/references/core-discipline.md +51 -0
- package/tdd/references/integration-testing.md +13 -0
- package/tdd/references/quality-model.md +68 -0
- package/tdd/references/unit-testing.md +15 -0
- package/tdd/schemas/quality-evidence.schema.json +211 -0
- package/tdd/schemas/quality-report.schema.json +110 -0
- package/tdd/scripts/quality-gate.js +475 -0
- package/to-spec/SKILL.md +40 -0
- package/to-spec/references/process.md +77 -0
- package/to-spec/scripts/check-project-docs.js +239 -0
- package/to-spec/scripts/project-docs-resolver.js +111 -0
- package/to-spec/templates/cli-reference.md +31 -0
- package/to-spec/templates/dev-doc.md +27 -0
- package/to-spec/templates/feature-spec.md +59 -0
- package/to-spec/templates/schema-doc.md +31 -0
- package/to-tickets/SKILL.md +42 -0
- package/to-tickets/references/publishing-and-templates.md +97 -0
- package/todo-driven-workflow/SKILL.md +28 -40
- package/todo-driven-workflow/references/execution-guide.md +27 -0
- package/using-git-worktrees/SKILL.md +29 -141
- package/using-git-worktrees/references/workflow-details.md +108 -0
- package/verification-loop/SKILL.md +29 -121
- package/verification-loop/references/verification-phases.md +116 -0
- package/verification-loop/templates/verification-report.template.md +15 -0
- package/verify-before-claim/SKILL.md +27 -70
- package/verify-before-claim/references/verification-guide.md +73 -0
- package/zoom-out/SKILL.md +28 -76
- package/zoom-out/references/circuit-breaker.md +66 -0
- package/zoom-out/templates/zoom-out-report.template.md +20 -0
- package/build-multi-agent-system/AGENTS.md +0 -50
- package/build-multi-agent-system/README.md +0 -63
- package/build-multi-agent-system/SKILL.md +0 -68
- package/create-agent-launcher/SKILL.md +0 -48
- package/create-agent-launcher/guidelines/platform-gemini.md +0 -101
- package/create-agent-launcher/workflows/01-init.md +0 -22
- package/create-agent-launcher/workflows/02-analysis.md +0 -26
- package/create-agent-launcher/workflows/03-generation.md +0 -29
- package/create-agent-launcher/workflows/04-launcher.md +0 -20
- package/docs/workflows/build-multi-agent-system.md +0 -62
- package/docs/workflows/create-agent-launcher.md +0 -62
- package/eval-framework/behavioral-test.js +0 -105
- package/eval-framework/cases/case2.json +0 -5
- package/eval-framework/mechanism-test.js +0 -251
- package/harness-everything/scripts/todo-cli.js +0 -157
- package/skill-style/STYLE.md +0 -37
|
@@ -0,0 +1,461 @@
|
|
|
1
|
+
const fs = require('fs');
|
|
2
|
+
const path = require('path');
|
|
3
|
+
const crypto = require('crypto');
|
|
4
|
+
const { pathToFileURL } = require('url');
|
|
5
|
+
const helper = require('./test-helper');
|
|
6
|
+
|
|
7
|
+
console.log('\n[2n] opencode enforcement plugin: real hook API and firing behavior...');
|
|
8
|
+
|
|
9
|
+
const pluginDir = path.join(helper.root, 'opencode-plugin');
|
|
10
|
+
const pluginFile = path.join(pluginDir, 'index.mjs');
|
|
11
|
+
const pluginManifestFile = path.join(pluginDir, 'plugin.json');
|
|
12
|
+
|
|
13
|
+
helper.check(
|
|
14
|
+
'2n. the repository inventory manifest exists',
|
|
15
|
+
fs.existsSync(pluginManifestFile),
|
|
16
|
+
`${pluginManifestFile} missing`
|
|
17
|
+
);
|
|
18
|
+
|
|
19
|
+
if (fs.existsSync(pluginManifestFile)) {
|
|
20
|
+
const manifest = JSON.parse(fs.readFileSync(pluginManifestFile, 'utf8'));
|
|
21
|
+
const onDisk = fs.readdirSync(path.join(pluginDir, 'hooks'))
|
|
22
|
+
.filter((file) => file.endsWith('.js'))
|
|
23
|
+
.sort();
|
|
24
|
+
const declared = Object.values(manifest.hooks || {})
|
|
25
|
+
.map((file) => path.relative(path.join(pluginDir, 'hooks'), path.join(pluginDir, file)))
|
|
26
|
+
.sort();
|
|
27
|
+
helper.check(
|
|
28
|
+
'2n. every standalone hook file on disk is declared in the inventory manifest',
|
|
29
|
+
onDisk.every((file) => declared.includes(file)),
|
|
30
|
+
`undeclared: ${onDisk.filter((file) => !declared.includes(file)).join(', ')}`
|
|
31
|
+
);
|
|
32
|
+
helper.check(
|
|
33
|
+
'2n. every declared inventory hook path exists on disk',
|
|
34
|
+
declared.every((file) => onDisk.includes(file)),
|
|
35
|
+
`dangling: ${declared.filter((file) => !onDisk.includes(file)).join(', ')}`
|
|
36
|
+
);
|
|
37
|
+
helper.check(
|
|
38
|
+
'2n. no standalone hook is declared more than once',
|
|
39
|
+
new Set(declared).size === declared.length,
|
|
40
|
+
declared.join(', ')
|
|
41
|
+
);
|
|
42
|
+
}
|
|
43
|
+
|
|
44
|
+
helper.check(
|
|
45
|
+
'2n. the plugin ships as a single self-contained module (installable by copying one file)',
|
|
46
|
+
fs.existsSync(pluginFile) && !/require\(['"]\.\//.test(fs.readFileSync(pluginFile, 'utf8')),
|
|
47
|
+
`${pluginFile} missing or imports a sibling file`
|
|
48
|
+
);
|
|
49
|
+
|
|
50
|
+
// Hook firing, against opencode's real Hooks interface (verified from
|
|
51
|
+
// packages/plugin/src/index.ts and packages/web/.../plugins.mdx in the
|
|
52
|
+
// opencode source - see opencode-plugin/README.md). Two isolations are
|
|
53
|
+
// load-bearing: HOME -> temp, because the plugin persists under
|
|
54
|
+
// ~/.agents/harness-everything/ and an un-redirected run would corrupt a
|
|
55
|
+
// real session's circuit-breaker state; workspace -> temp, with a
|
|
56
|
+
// package.json this test controls, so verification pass/fail is
|
|
57
|
+
// deterministic.
|
|
58
|
+
(async () => {
|
|
59
|
+
const fakeHome = helper.tempDir('.mechanism-test-opencode-home');
|
|
60
|
+
const workspace = helper.tempDir('.mechanism-test-opencode-workspace');
|
|
61
|
+
fs.mkdirSync(fakeHome, { recursive: true });
|
|
62
|
+
fs.mkdirSync(workspace, { recursive: true });
|
|
63
|
+
process.env.HOME = fakeHome;
|
|
64
|
+
process.env.USERPROFILE = fakeHome;
|
|
65
|
+
delete process.env.HARNESS_STATE_HOME;
|
|
66
|
+
|
|
67
|
+
// Mirrors the plugin's own getWorkspaceKey(directory) exactly (issue #42
|
|
68
|
+
// item #4) - duplicated here rather than imported since the plugin only
|
|
69
|
+
// exports HarnessEnforcement.
|
|
70
|
+
const realWorkspace = fs.realpathSync(path.resolve(workspace));
|
|
71
|
+
const slug = path.basename(realWorkspace).toLowerCase().replace(/[^a-z0-9._-]+/g, '-').replace(/^-+|-+$/g, '') || 'workspace';
|
|
72
|
+
const hashInput = process.platform === 'win32' ? realWorkspace.toLowerCase() : realWorkspace;
|
|
73
|
+
const hash = crypto.createHash('sha1').update(hashInput).digest('hex').slice(0, 12);
|
|
74
|
+
const stateHome = path.join(fakeHome, '.agents', 'harness-everything');
|
|
75
|
+
const stateRoot = path.join(stateHome, 'workspaces', `${slug}-${hash}`, 'state');
|
|
76
|
+
const stateDir = path.join(stateRoot, 'sessions', 's1');
|
|
77
|
+
const secondSessionDir = path.join(stateRoot, 'sessions', 's2');
|
|
78
|
+
const retrySessionDir = path.join(stateRoot, 'sessions', 'prompt-retry');
|
|
79
|
+
const editStateFile = path.join(stateDir, 'edit-state.json');
|
|
80
|
+
const breakerFile = path.join(stateDir, 'circuit-breaker.json');
|
|
81
|
+
const complianceFile = path.join(stateDir, 'compliance.json');
|
|
82
|
+
const legacyDir = path.join(fakeHome, '.harness-state');
|
|
83
|
+
fs.mkdirSync(legacyDir, { recursive: true });
|
|
84
|
+
fs.writeFileSync(path.join(legacyDir, 'circuit-breaker.json'), JSON.stringify({ legacy: true }));
|
|
85
|
+
|
|
86
|
+
function writePackageScript(script) {
|
|
87
|
+
fs.writeFileSync(
|
|
88
|
+
path.join(workspace, 'package.json'),
|
|
89
|
+
JSON.stringify({ name: 'fixture', version: '0.0.0', scripts: { test: script } }, null, 2)
|
|
90
|
+
);
|
|
91
|
+
}
|
|
92
|
+
|
|
93
|
+
const promptCalls = [];
|
|
94
|
+
let rejectNextPrompt = false;
|
|
95
|
+
const mockClient = {
|
|
96
|
+
session: {
|
|
97
|
+
prompt: async (opts) => {
|
|
98
|
+
promptCalls.push(opts);
|
|
99
|
+
if (rejectNextPrompt) {
|
|
100
|
+
rejectNextPrompt = false;
|
|
101
|
+
throw new Error('prompt transport unavailable');
|
|
102
|
+
}
|
|
103
|
+
return {};
|
|
104
|
+
}
|
|
105
|
+
}
|
|
106
|
+
};
|
|
107
|
+
|
|
108
|
+
const { HarnessEnforcement } = await import(pathToFileURL(pluginFile).href);
|
|
109
|
+
|
|
110
|
+
const hooks = await HarnessEnforcement({
|
|
111
|
+
client: mockClient,
|
|
112
|
+
directory: workspace,
|
|
113
|
+
worktree: workspace,
|
|
114
|
+
project: {},
|
|
115
|
+
serverUrl: new URL('http://localhost:4096'),
|
|
116
|
+
experimental_workspace: { register() {} },
|
|
117
|
+
$: async () => {}
|
|
118
|
+
});
|
|
119
|
+
|
|
120
|
+
helper.check(
|
|
121
|
+
'2n. exports exactly the hooks it uses (tool.execute.before/after, event)',
|
|
122
|
+
['tool.execute.before', 'tool.execute.after', 'event'].every((k) => typeof hooks[k] === 'function') &&
|
|
123
|
+
Object.keys(hooks).length === 3,
|
|
124
|
+
`got keys: ${Object.keys(hooks).join(', ')}`
|
|
125
|
+
);
|
|
126
|
+
|
|
127
|
+
// --- edit tracking (tool.execute.after) ---
|
|
128
|
+
await hooks['tool.execute.after']({ tool: 'read', sessionID: 's1', callID: 'c0' }, { title: '', output: '', metadata: {} });
|
|
129
|
+
helper.check('2n. non-edit tools do not mark verification pending', !fs.existsSync(editStateFile), 'edit-state.json created by a read');
|
|
130
|
+
|
|
131
|
+
await hooks['tool.execute.after']({ tool: 'edit', sessionID: 's1', callID: 'c1' }, { title: '', output: '', metadata: {} });
|
|
132
|
+
let editState = JSON.parse(fs.readFileSync(editStateFile, 'utf8'));
|
|
133
|
+
helper.check(
|
|
134
|
+
'2n. an edit tool call marks verification pending',
|
|
135
|
+
editState.verificationPending === true && editState.editsSinceVerification === 1,
|
|
136
|
+
JSON.stringify(editState)
|
|
137
|
+
);
|
|
138
|
+
|
|
139
|
+
// --- verification gate on session.idle: passing case clears pending ---
|
|
140
|
+
writePackageScript('node -e "process.exit(0)"');
|
|
141
|
+
await hooks.event({ event: { type: 'session.idle', properties: { sessionID: 's1' } } });
|
|
142
|
+
editState = JSON.parse(fs.readFileSync(editStateFile, 'utf8'));
|
|
143
|
+
helper.check(
|
|
144
|
+
'2n. a passing verification on session.idle clears pending state',
|
|
145
|
+
editState.verificationPending === false && editState.editsSinceVerification === 0,
|
|
146
|
+
JSON.stringify(editState)
|
|
147
|
+
);
|
|
148
|
+
helper.check('2n. no follow-up prompt is sent when verification passes', promptCalls.length === 0, `${promptCalls.length} prompt(s) sent`);
|
|
149
|
+
|
|
150
|
+
// --- verification gate: failing case drives the circuit breaker ---
|
|
151
|
+
writePackageScript('node -e "process.exit(1)"');
|
|
152
|
+
|
|
153
|
+
for (let i = 1; i <= 2; i++) {
|
|
154
|
+
await hooks['tool.execute.after']({ tool: 'edit', sessionID: 's1', callID: `c${i + 1}` }, { title: '', output: '', metadata: {} });
|
|
155
|
+
await hooks.event({ event: { type: 'session.idle', properties: { sessionID: 's1' } } });
|
|
156
|
+
}
|
|
157
|
+
let breaker = JSON.parse(fs.readFileSync(breakerFile, 'utf8'));
|
|
158
|
+
const sig = Object.keys(breaker.failures)[0];
|
|
159
|
+
helper.check(
|
|
160
|
+
'2n. two failures on the same signature: still allowed, not yet tripped',
|
|
161
|
+
breaker.failures[sig].count === 2 && breaker.hardLock === false,
|
|
162
|
+
JSON.stringify(breaker)
|
|
163
|
+
);
|
|
164
|
+
helper.check('2n. each failed verification sends exactly one follow-up prompt', promptCalls.length === 2, `${promptCalls.length} prompt(s) sent`);
|
|
165
|
+
|
|
166
|
+
// 3rd failure -> force reflection
|
|
167
|
+
await hooks['tool.execute.after']({ tool: 'edit', sessionID: 's1', callID: 'c4' }, { title: '', output: '', metadata: {} });
|
|
168
|
+
await hooks.event({ event: { type: 'session.idle', properties: { sessionID: 's1' } } });
|
|
169
|
+
breaker = JSON.parse(fs.readFileSync(breakerFile, 'utf8'));
|
|
170
|
+
let compliance = JSON.parse(fs.readFileSync(complianceFile, 'utf8'));
|
|
171
|
+
helper.check(
|
|
172
|
+
'2n. 3rd failure on the same signature forces reflection',
|
|
173
|
+
breaker.failures[sig].count === 3 && compliance.reflectionsForced === 1,
|
|
174
|
+
JSON.stringify({ breaker, compliance })
|
|
175
|
+
);
|
|
176
|
+
helper.check(
|
|
177
|
+
'2n. the forced-reflection prompt says to reflect',
|
|
178
|
+
/reflect/i.test(promptCalls[promptCalls.length - 1].body.parts[0].text),
|
|
179
|
+
promptCalls[promptCalls.length - 1].body.parts[0].text
|
|
180
|
+
);
|
|
181
|
+
|
|
182
|
+
let pendingEditBlocked = false;
|
|
183
|
+
try {
|
|
184
|
+
await hooks['tool.execute.before']({ tool: 'edit', sessionID: 's1', callID: 'pending-edit' }, { args: {} });
|
|
185
|
+
} catch {
|
|
186
|
+
pendingEditBlocked = true;
|
|
187
|
+
}
|
|
188
|
+
helper.check('2n. pending reflection blocks another code edit', pendingEditBlocked, 'tool.execute.before allowed an edit before the report');
|
|
189
|
+
|
|
190
|
+
let pendingReadAllowed = true;
|
|
191
|
+
try {
|
|
192
|
+
await hooks['tool.execute.before']({ tool: 'read', sessionID: 's1', callID: 'pending-read' }, { args: {} });
|
|
193
|
+
} catch {
|
|
194
|
+
pendingReadAllowed = false;
|
|
195
|
+
}
|
|
196
|
+
helper.check('2n. pending reflection still allows read tools', pendingReadAllowed, 'tool.execute.before blocked a read during reflection');
|
|
197
|
+
|
|
198
|
+
// Complete the reflection through the real hook sequence: the forced
|
|
199
|
+
// prompt names a token and artifact, then the agent writes that artifact
|
|
200
|
+
// through an edit-shaped tool call. No test seeding of breaker state.
|
|
201
|
+
const reflectionPrompt = promptCalls[promptCalls.length - 1].body.parts[0].text;
|
|
202
|
+
const reflectionTokenMatch = reflectionPrompt.match(/reflection token: ([A-Za-z0-9_-]+)/i);
|
|
203
|
+
helper.check('2n. the forced-reflection prompt names a verifiable artifact token', !!reflectionTokenMatch, reflectionPrompt);
|
|
204
|
+
const reflectionToken = reflectionTokenMatch && reflectionTokenMatch[1];
|
|
205
|
+
const reflectionFile = path.join(stateDir, 'zoom-out-report.md');
|
|
206
|
+
fs.writeFileSync(
|
|
207
|
+
reflectionFile,
|
|
208
|
+
`RESUME: misplaced\n## Goal\nfix the fixture\n## Failed Attempts\nthree retries\n## Verified Facts\ntest remains red\n## Diagnosis\nthe fixture intentionally fails\n## Decision\nA decision without a directive\nReflection token: ${reflectionToken}\n`,
|
|
209
|
+
'utf8'
|
|
210
|
+
);
|
|
211
|
+
await hooks['tool.execute.after'](
|
|
212
|
+
{ tool: 'write', sessionID: 's1', callID: 'reflection' },
|
|
213
|
+
{ title: '', output: '', metadata: {}, args: { filePath: reflectionFile } }
|
|
214
|
+
);
|
|
215
|
+
breaker = JSON.parse(fs.readFileSync(breakerFile, 'utf8'));
|
|
216
|
+
helper.check(
|
|
217
|
+
'2n. a directive outside the Decision section does not complete reflection',
|
|
218
|
+
breaker.reflectionPending === true && breaker.lastReflection === null,
|
|
219
|
+
JSON.stringify(breaker)
|
|
220
|
+
);
|
|
221
|
+
|
|
222
|
+
fs.writeFileSync(
|
|
223
|
+
reflectionFile,
|
|
224
|
+
`## Goal\nfix the fixture\n## Failed Attempts\nthree retries\n## Verified Facts\ntest remains red\n## Diagnosis\nthe fixture intentionally fails\n## Decision\nRESUME: change the fixture\nReflection token: ${reflectionToken}\n`,
|
|
225
|
+
'utf8'
|
|
226
|
+
);
|
|
227
|
+
await hooks['tool.execute.after'](
|
|
228
|
+
{ tool: 'apply_patch', sessionID: 's1', callID: 'reflection-patch' },
|
|
229
|
+
{
|
|
230
|
+
title: '',
|
|
231
|
+
output: '',
|
|
232
|
+
metadata: {},
|
|
233
|
+
args: { patchText: `*** Begin Patch\n*** Update File: ${reflectionFile}\n@@\n*** End Patch` }
|
|
234
|
+
}
|
|
235
|
+
);
|
|
236
|
+
breaker = JSON.parse(fs.readFileSync(breakerFile, 'utf8'));
|
|
237
|
+
helper.check(
|
|
238
|
+
'2n. apply_patch reflection completes the lifecycle and persists lastReflection',
|
|
239
|
+
Number.isFinite(breaker.lastReflection) && breaker.lastReflection > 0 && breaker.reflectionPending === false,
|
|
240
|
+
JSON.stringify(breaker)
|
|
241
|
+
);
|
|
242
|
+
|
|
243
|
+
// A repeated idle event after the follow-up is already pending is a no-op.
|
|
244
|
+
// The breaker must wait for a fresh code edit before counting another try.
|
|
245
|
+
const promptsAfterReflection = promptCalls.length;
|
|
246
|
+
await hooks.event({ event: { type: 'session.idle', properties: { sessionID: 's1' } } });
|
|
247
|
+
helper.check(
|
|
248
|
+
'2n. repeated idle does not duplicate verification or follow-up prompts',
|
|
249
|
+
promptCalls.length === promptsAfterReflection,
|
|
250
|
+
`${promptCalls.length} prompt(s) sent`
|
|
251
|
+
);
|
|
252
|
+
|
|
253
|
+
const patchSessionDir = path.join(stateRoot, 'sessions', 'patch-only');
|
|
254
|
+
const patchBreakerFile = path.join(patchSessionDir, 'circuit-breaker.json');
|
|
255
|
+
const patchReflectionFile = path.join(patchSessionDir, 'zoom-out-report.md');
|
|
256
|
+
const patchReflectionToken = 'patch-reflection-token';
|
|
257
|
+
fs.mkdirSync(patchSessionDir, { recursive: true });
|
|
258
|
+
fs.writeFileSync(
|
|
259
|
+
patchBreakerFile,
|
|
260
|
+
JSON.stringify({
|
|
261
|
+
failures: { fixture: { count: 3, firstSeen: Date.now() - 1000 } },
|
|
262
|
+
hardLock: false,
|
|
263
|
+
lastReflection: null,
|
|
264
|
+
lastReflectionSignature: null,
|
|
265
|
+
reflectionPending: true,
|
|
266
|
+
reflectionRequestedAt: Date.now(),
|
|
267
|
+
reflectionToken: patchReflectionToken,
|
|
268
|
+
reflectionSignature: 'fixture'
|
|
269
|
+
}, null, 2),
|
|
270
|
+
'utf8'
|
|
271
|
+
);
|
|
272
|
+
fs.writeFileSync(
|
|
273
|
+
patchReflectionFile,
|
|
274
|
+
`## Goal\nfix the fixture\n## Failed Attempts\nthree retries\n## Verified Facts\ntest remains red\n## Diagnosis\nthe fixture intentionally fails\n## Decision\nRESUME: change the fixture\nReflection token: ${patchReflectionToken}\n`,
|
|
275
|
+
'utf8'
|
|
276
|
+
);
|
|
277
|
+
const patchText = `*** Begin Patch\n*** Update File: ${patchReflectionFile}\n@@\n*** End Patch`;
|
|
278
|
+
let patchBeforeAllowed = true;
|
|
279
|
+
try {
|
|
280
|
+
await hooks['tool.execute.before'](
|
|
281
|
+
{ tool: 'apply_patch', sessionID: 'patch-only', callID: 'patch-before' },
|
|
282
|
+
{ args: { patchText } }
|
|
283
|
+
);
|
|
284
|
+
} catch (error) {
|
|
285
|
+
patchBeforeAllowed = false;
|
|
286
|
+
}
|
|
287
|
+
helper.check(
|
|
288
|
+
'2n. apply_patch is allowed to write a pending reflection artifact',
|
|
289
|
+
patchBeforeAllowed,
|
|
290
|
+
'tool.execute.before blocked output.args.patchText'
|
|
291
|
+
);
|
|
292
|
+
await hooks['tool.execute.after'](
|
|
293
|
+
{ tool: 'apply_patch', sessionID: 'patch-only', callID: 'patch-after' },
|
|
294
|
+
{ title: '', output: '', metadata: {}, args: { patchText } }
|
|
295
|
+
);
|
|
296
|
+
const patchBreaker = JSON.parse(fs.readFileSync(patchBreakerFile, 'utf8'));
|
|
297
|
+
helper.check(
|
|
298
|
+
'2n. apply_patch output args complete a pending reflection',
|
|
299
|
+
patchBreaker.reflectionPending === false && Number.isFinite(patchBreaker.lastReflection),
|
|
300
|
+
JSON.stringify(patchBreaker)
|
|
301
|
+
);
|
|
302
|
+
|
|
303
|
+
const mixedPatchSessionDir = path.join(stateRoot, 'sessions', 'mixed-patch');
|
|
304
|
+
const mixedPatchBreakerFile = path.join(mixedPatchSessionDir, 'circuit-breaker.json');
|
|
305
|
+
const mixedPatchReflectionFile = path.join(mixedPatchSessionDir, 'zoom-out-report.md');
|
|
306
|
+
fs.mkdirSync(mixedPatchSessionDir, { recursive: true });
|
|
307
|
+
fs.writeFileSync(mixedPatchBreakerFile, JSON.stringify({
|
|
308
|
+
failures: { fixture: { count: 3, firstSeen: Date.now() - 1000 } },
|
|
309
|
+
hardLock: false,
|
|
310
|
+
lastReflection: null,
|
|
311
|
+
lastReflectionSignature: null,
|
|
312
|
+
reflectionPending: true,
|
|
313
|
+
reflectionRequestedAt: Date.now(),
|
|
314
|
+
reflectionToken: 'mixed-patch-token',
|
|
315
|
+
reflectionSignature: 'fixture'
|
|
316
|
+
}, null, 2));
|
|
317
|
+
fs.writeFileSync(mixedPatchReflectionFile, [
|
|
318
|
+
'## Goal', 'fix the fixture', '## Failed Attempts', 'three retries',
|
|
319
|
+
'## Verified Facts', 'test remains red', '## Diagnosis', 'the fixture intentionally fails',
|
|
320
|
+
'## Decision', 'RESUME: change the fixture', 'Reflection token: mixed-patch-token', ''
|
|
321
|
+
].join('\n'));
|
|
322
|
+
let mixedPatchBlocked = false;
|
|
323
|
+
try {
|
|
324
|
+
await hooks['tool.execute.before'](
|
|
325
|
+
{ tool: 'apply_patch', sessionID: 'mixed-patch', callID: 'mixed-before' },
|
|
326
|
+
{ args: { patchText: `*** Begin Patch\n*** Update File: ${mixedPatchReflectionFile}\n@@\n*** Update File: ${path.join(workspace, 'unrelated.js')}\n@@\n*** End Patch` } }
|
|
327
|
+
);
|
|
328
|
+
} catch {
|
|
329
|
+
mixedPatchBlocked = true;
|
|
330
|
+
}
|
|
331
|
+
helper.check('2n. mixed apply_patch cannot bypass a pending reflection', mixedPatchBlocked, 'unrelated files were accepted with the reflection artifact');
|
|
332
|
+
|
|
333
|
+
// A new code edit after a completed reflection is the post-reflection retry.
|
|
334
|
+
await hooks['tool.execute.after']({ tool: 'edit', sessionID: 's1', callID: 'c5' }, { title: '', output: '', metadata: {} });
|
|
335
|
+
await hooks.event({ event: { type: 'session.idle', properties: { sessionID: 's1' } } });
|
|
336
|
+
breaker = JSON.parse(fs.readFileSync(breakerFile, 'utf8'));
|
|
337
|
+
helper.check('2n. the same failure returning after a reflection hard-locks the breaker', breaker.hardLock === true, JSON.stringify(breaker));
|
|
338
|
+
helper.check(
|
|
339
|
+
'2n. the hard-lock prompt says locked',
|
|
340
|
+
/lock/i.test(promptCalls[promptCalls.length - 1].body.parts[0].text),
|
|
341
|
+
promptCalls[promptCalls.length - 1].body.parts[0].text
|
|
342
|
+
);
|
|
343
|
+
|
|
344
|
+
// --- circuit breaker enforcement (tool.execute.before) ---
|
|
345
|
+
let blocked = false;
|
|
346
|
+
try {
|
|
347
|
+
await hooks['tool.execute.before']({ tool: 'edit', sessionID: 's1', callID: 'c6' }, { args: {} });
|
|
348
|
+
} catch {
|
|
349
|
+
blocked = true;
|
|
350
|
+
}
|
|
351
|
+
helper.check('2n. hard-locked breaker throws on the next edit attempt', blocked, 'tool.execute.before did not throw');
|
|
352
|
+
|
|
353
|
+
let readBlocked = false;
|
|
354
|
+
try {
|
|
355
|
+
await hooks['tool.execute.before']({ tool: 'read', sessionID: 's1', callID: 'c7' }, { args: {} });
|
|
356
|
+
} catch {
|
|
357
|
+
readBlocked = true;
|
|
358
|
+
}
|
|
359
|
+
helper.check('2n. a hard lock only blocks edit-shaped tools, not read', !readBlocked, 'tool.execute.before blocked a read');
|
|
360
|
+
|
|
361
|
+
// The same workspace can host independent sessions. Session s2 starts with
|
|
362
|
+
// a fresh breaker and cannot inherit s1's hard lock.
|
|
363
|
+
await hooks['tool.execute.after']({ tool: 'edit', sessionID: 's2', callID: 's2-c1', args: {} }, { title: '', output: '', metadata: {} });
|
|
364
|
+
await hooks.event({ event: { type: 'session.idle', properties: { sessionID: 's2' } } });
|
|
365
|
+
const secondBreaker = JSON.parse(fs.readFileSync(path.join(secondSessionDir, 'circuit-breaker.json'), 'utf8'));
|
|
366
|
+
helper.check(
|
|
367
|
+
'2n. a second session has an independent breaker stream',
|
|
368
|
+
secondBreaker.hardLock === false && Object.values(secondBreaker.failures).some((entry) => entry.count === 1),
|
|
369
|
+
JSON.stringify(secondBreaker)
|
|
370
|
+
);
|
|
371
|
+
|
|
372
|
+
const corruptSessionDir = path.join(stateRoot, 'sessions', 'corrupt');
|
|
373
|
+
const corruptBreakerFile = path.join(corruptSessionDir, 'circuit-breaker.json');
|
|
374
|
+
fs.mkdirSync(corruptSessionDir, { recursive: true });
|
|
375
|
+
fs.writeFileSync(corruptBreakerFile, '{not-json', 'utf8');
|
|
376
|
+
let corruptStateBlocked = false;
|
|
377
|
+
let corruptStateMessage = '';
|
|
378
|
+
try {
|
|
379
|
+
await hooks['tool.execute.before']({ tool: 'edit', sessionID: 'corrupt', callID: 'corrupt-edit' }, { args: {} });
|
|
380
|
+
} catch (error) {
|
|
381
|
+
corruptStateBlocked = true;
|
|
382
|
+
corruptStateMessage = String(error && error.message ? error.message : error);
|
|
383
|
+
}
|
|
384
|
+
helper.check(
|
|
385
|
+
'2n. corrupt breaker state fails closed before an edit',
|
|
386
|
+
corruptStateBlocked && /corrupt/i.test(corruptStateMessage),
|
|
387
|
+
corruptStateMessage || 'tool.execute.before allowed an edit with corrupt breaker state'
|
|
388
|
+
);
|
|
389
|
+
|
|
390
|
+
// A prompt transport failure is retryable without re-running verification
|
|
391
|
+
// or incrementing the circuit-breaker count.
|
|
392
|
+
const promptsBeforeDeliveryFailure = promptCalls.length;
|
|
393
|
+
rejectNextPrompt = true;
|
|
394
|
+
await hooks['tool.execute.after']({ tool: 'edit', sessionID: 'prompt-retry', callID: 'retry-c1' }, { title: '', output: '', metadata: {} });
|
|
395
|
+
await hooks.event({ event: { type: 'session.idle', properties: { sessionID: 'prompt-retry' } } });
|
|
396
|
+
let retryEditState = JSON.parse(fs.readFileSync(path.join(retrySessionDir, 'edit-state.json'), 'utf8'));
|
|
397
|
+
let retryBreaker = JSON.parse(fs.readFileSync(path.join(retrySessionDir, 'circuit-breaker.json'), 'utf8'));
|
|
398
|
+
helper.check(
|
|
399
|
+
'2n. failed prompt delivery remains retryable',
|
|
400
|
+
promptCalls.length === promptsBeforeDeliveryFailure + 1 &&
|
|
401
|
+
retryEditState.followUpPending === false &&
|
|
402
|
+
retryEditState.followUpDeliveryPending === true &&
|
|
403
|
+
retryBreaker.failures[Object.keys(retryBreaker.failures)[0]].count === 1,
|
|
404
|
+
JSON.stringify({ retryEditState, retryBreaker })
|
|
405
|
+
);
|
|
406
|
+
await hooks.event({ event: { type: 'session.idle', properties: { sessionID: 'prompt-retry' } } });
|
|
407
|
+
retryEditState = JSON.parse(fs.readFileSync(path.join(retrySessionDir, 'edit-state.json'), 'utf8'));
|
|
408
|
+
retryBreaker = JSON.parse(fs.readFileSync(path.join(retrySessionDir, 'circuit-breaker.json'), 'utf8'));
|
|
409
|
+
helper.check(
|
|
410
|
+
'2n. next idle retries the saved follow-up exactly once',
|
|
411
|
+
promptCalls.length === promptsBeforeDeliveryFailure + 2 &&
|
|
412
|
+
retryEditState.followUpPending === true &&
|
|
413
|
+
retryEditState.followUpDeliveryPending === false &&
|
|
414
|
+
retryBreaker.failures[Object.keys(retryBreaker.failures)[0]].count === 1,
|
|
415
|
+
JSON.stringify({ retryEditState, retryBreaker })
|
|
416
|
+
);
|
|
417
|
+
|
|
418
|
+
// A fresh session event is the explicit reset contract for a reused session
|
|
419
|
+
// identity, and it removes only that session's plugin state.
|
|
420
|
+
await hooks.event({ event: { type: 'session.created', properties: { info: { id: 's1' } } } });
|
|
421
|
+
helper.check('2n. session.created resets only the current session state', !fs.existsSync(breakerFile), breakerFile);
|
|
422
|
+
let resetAllowed = true;
|
|
423
|
+
try {
|
|
424
|
+
await hooks['tool.execute.before']({ tool: 'edit', sessionID: 's1', callID: 'after-reset' }, { args: {} });
|
|
425
|
+
} catch {
|
|
426
|
+
resetAllowed = false;
|
|
427
|
+
}
|
|
428
|
+
helper.check('2n. a reset session can edit again', resetAllowed, 'tool.execute.before remained locked');
|
|
429
|
+
|
|
430
|
+
// Invalid session IDs are hashed into a child directory. Resetting one can
|
|
431
|
+
// therefore never recurse into the workspace state parent or another path.
|
|
432
|
+
const sessionsRoot = path.join(stateRoot, 'sessions');
|
|
433
|
+
const outsideSentinel = path.join(stateHome, 'outside-session-sentinel');
|
|
434
|
+
fs.mkdirSync(outsideSentinel, { recursive: true });
|
|
435
|
+
fs.writeFileSync(path.join(outsideSentinel, 'keep.txt'), 'keep');
|
|
436
|
+
await hooks['tool.execute.after']({ tool: 'edit', sessionID: '..', callID: 'dotdot' }, { title: '', output: '', metadata: {} });
|
|
437
|
+
await hooks['tool.execute.after']({ tool: 'edit', sessionID: '.', callID: 'dot' }, { title: '', output: '', metadata: {} });
|
|
438
|
+
await hooks.event({ event: { type: 'session.created', properties: { info: { id: '../../outside-session-sentinel' } } } });
|
|
439
|
+
const sessionEntries = fs.readdirSync(sessionsRoot, { withFileTypes: true });
|
|
440
|
+
helper.check(
|
|
441
|
+
'2n. dot and dotdot session IDs never escape the sessions root',
|
|
442
|
+
!fs.existsSync(path.join(stateRoot, 'edit-state.json')) &&
|
|
443
|
+
sessionEntries.every((entry) => entry.isDirectory() && !['.', '..'].includes(entry.name)),
|
|
444
|
+
sessionEntries.map((entry) => entry.name).join(', ')
|
|
445
|
+
);
|
|
446
|
+
helper.check('2n. resetting an unsafe session cannot delete an outside sentinel', fs.existsSync(path.join(outsideSentinel, 'keep.txt')), outsideSentinel);
|
|
447
|
+
helper.check('2n. resetting s1 leaves the independent s2 state', fs.existsSync(path.join(secondSessionDir, 'circuit-breaker.json')), secondSessionDir);
|
|
448
|
+
|
|
449
|
+
const workspaceStateRoot = path.dirname(stateRoot);
|
|
450
|
+
helper.check(
|
|
451
|
+
'2n. legacy flat state migrates into the workspace-keyed root',
|
|
452
|
+
!fs.existsSync(legacyDir) && fs.existsSync(path.join(workspaceStateRoot, 'circuit-breaker.json')),
|
|
453
|
+
JSON.stringify({ legacyDir, workspaceStateRoot })
|
|
454
|
+
);
|
|
455
|
+
helper.check('2n. hook state stayed inside the redirected HOME', fs.existsSync(stateRoot), `no state root under ${stateHome}`);
|
|
456
|
+
|
|
457
|
+
helper.finish();
|
|
458
|
+
})().catch((err) => {
|
|
459
|
+
console.error('2n. opencode plugin test crashed:', err);
|
|
460
|
+
process.exit(1);
|
|
461
|
+
});
|
|
@@ -0,0 +1,183 @@
|
|
|
1
|
+
const fs = require('fs');
|
|
2
|
+
const os = require('os');
|
|
3
|
+
const path = require('path');
|
|
4
|
+
const { spawnSync } = require('child_process');
|
|
5
|
+
const helper = require('./test-helper');
|
|
6
|
+
const { getRuntimeRoot } = require('../multi-agent-workspace/scripts/project-docs-resolver');
|
|
7
|
+
|
|
8
|
+
console.log('\n[2o] State-home isolation: runtime state never scatters into cwd (issue #42)...');
|
|
9
|
+
|
|
10
|
+
const repoRoot = helper.root;
|
|
11
|
+
let failures = 0;
|
|
12
|
+
function check(name, condition, detail) {
|
|
13
|
+
if (condition) {
|
|
14
|
+
console.log(`✅ ${name}`);
|
|
15
|
+
} else {
|
|
16
|
+
console.error(`❌ ${name}`);
|
|
17
|
+
if (detail) console.error(` ${detail}`);
|
|
18
|
+
failures++;
|
|
19
|
+
}
|
|
20
|
+
}
|
|
21
|
+
|
|
22
|
+
function canonicalPath(value) {
|
|
23
|
+
const resolved = path.resolve(value);
|
|
24
|
+
try { return fs.realpathSync(resolved); } catch (e) { return resolved; }
|
|
25
|
+
}
|
|
26
|
+
|
|
27
|
+
// Everything below runs with HOME/USERPROFILE redirected to a throwaway
|
|
28
|
+
// fake home, so this test can never touch (or be polluted by) the real
|
|
29
|
+
// developer's ~/.agents/harness-everything - same isolation approach as
|
|
30
|
+
// mechanism-2n-opencode-plugin.test.js.
|
|
31
|
+
const fakeHome = fs.mkdtempSync(path.join(os.tmpdir(), 'harness-2o-home-'));
|
|
32
|
+
const stateHome = path.join(fakeHome, '.agents', 'harness-everything');
|
|
33
|
+
const baseEnv = { ...process.env, HOME: fakeHome, USERPROFILE: fakeHome };
|
|
34
|
+
delete baseEnv.HARNESS_STATE_HOME;
|
|
35
|
+
delete baseEnv.CLAUDE;
|
|
36
|
+
baseEnv.HARNESS_STATE_HOME = stateHome;
|
|
37
|
+
process.env.HARNESS_STATE_HOME = stateHome;
|
|
38
|
+
|
|
39
|
+
function listFilesRecursive(dir) {
|
|
40
|
+
if (!fs.existsSync(dir)) return [];
|
|
41
|
+
const out = [];
|
|
42
|
+
for (const entry of fs.readdirSync(dir, { withFileTypes: true })) {
|
|
43
|
+
const p = path.join(dir, entry.name);
|
|
44
|
+
if (entry.isDirectory()) out.push(...listFilesRecursive(p));
|
|
45
|
+
else out.push(p);
|
|
46
|
+
}
|
|
47
|
+
return out;
|
|
48
|
+
}
|
|
49
|
+
|
|
50
|
+
function listAllRecursive(dir) {
|
|
51
|
+
if (!fs.existsSync(dir)) return [];
|
|
52
|
+
const out = [];
|
|
53
|
+
for (const entry of fs.readdirSync(dir, { withFileTypes: true })) {
|
|
54
|
+
const p = path.join(dir, entry.name);
|
|
55
|
+
out.push(p);
|
|
56
|
+
if (entry.isDirectory()) out.push(...listAllRecursive(p));
|
|
57
|
+
}
|
|
58
|
+
return out;
|
|
59
|
+
}
|
|
60
|
+
|
|
61
|
+
function run(cwd, args, env) {
|
|
62
|
+
return spawnSync(process.execPath, args, { cwd, env: env || baseEnv, encoding: 'utf8' });
|
|
63
|
+
}
|
|
64
|
+
|
|
65
|
+
const noGitFixture = fs.mkdtempSync(path.join(os.tmpdir(), 'harness-2o-nogit-'));
|
|
66
|
+
const nestedNoGit = path.join(noGitFixture, 'fixtures', 'some-test-dir');
|
|
67
|
+
fs.mkdirSync(nestedNoGit, { recursive: true });
|
|
68
|
+
|
|
69
|
+
{
|
|
70
|
+
const inline = `
|
|
71
|
+
const { getSessionDir } = require(${JSON.stringify(path.join(repoRoot, 'hooks', 'scripts', 'lib', 'harness-state.js'))});
|
|
72
|
+
getSessionDir(undefined, 'issue42-repro-session');
|
|
73
|
+
`;
|
|
74
|
+
const res = run(nestedNoGit, ['-e', inline]);
|
|
75
|
+
check(
|
|
76
|
+
'2o. harness-state.js creates zero files in a non-git nested dir',
|
|
77
|
+
res.status === 0 && listFilesRecursive(nestedNoGit).length === 0,
|
|
78
|
+
`exit=${res.status} stderr=${res.stderr} files=${JSON.stringify(listFilesRecursive(nestedNoGit))}`
|
|
79
|
+
);
|
|
80
|
+
check(
|
|
81
|
+
'2o. ...and the session dir lands under the global state home instead',
|
|
82
|
+
fs.existsSync(path.join(stateHome, 'workspaces')) && listAllRecursive(stateHome).some(f => f.includes('issue42-repro-session')),
|
|
83
|
+
`nothing under ${stateHome}: ${JSON.stringify(listAllRecursive(stateHome))}`
|
|
84
|
+
);
|
|
85
|
+
}
|
|
86
|
+
|
|
87
|
+
{
|
|
88
|
+
const modelSelector = path.join(repoRoot, 'fable-mode', 'scripts', 'model-selector.js');
|
|
89
|
+
const res = run(nestedNoGit, [
|
|
90
|
+
modelSelector, '--requested', 'haiku', '--available', 'haiku,sonnet',
|
|
91
|
+
'--available-agents', 'fable-worker-haiku', '--stage-brief', 'x', '--pass-condition', 'y',
|
|
92
|
+
'--verification-command', 'npm test', '--verifier-result', 'pending'
|
|
93
|
+
]);
|
|
94
|
+
check(
|
|
95
|
+
'2o. model-selector.js (zero-walk before the fix) creates zero files in a non-git nested dir',
|
|
96
|
+
res.status === 0 && listFilesRecursive(nestedNoGit).length === 0,
|
|
97
|
+
`exit=${res.status} stderr=${res.stderr} files=${JSON.stringify(listFilesRecursive(nestedNoGit))}`
|
|
98
|
+
);
|
|
99
|
+
check(
|
|
100
|
+
'2o. ...and its audit log lands under the same global state/ stream as harness-state.js',
|
|
101
|
+
listFilesRecursive(stateHome).some(f => f.endsWith(path.join('fable-mode', 'audit.jsonl'))),
|
|
102
|
+
`no fable-mode/audit.jsonl under ${stateHome}`
|
|
103
|
+
);
|
|
104
|
+
}
|
|
105
|
+
|
|
106
|
+
fs.rmSync(noGitFixture, { recursive: true, force: true });
|
|
107
|
+
|
|
108
|
+
const gitFixture = fs.mkdtempSync(path.join(os.tmpdir(), 'harness-2o-git-'));
|
|
109
|
+
spawnSync('git', ['init', '--quiet'], { cwd: gitFixture });
|
|
110
|
+
const nestedInGit = path.join(gitFixture, 'packages', 'nested');
|
|
111
|
+
fs.mkdirSync(nestedInGit, { recursive: true });
|
|
112
|
+
|
|
113
|
+
{
|
|
114
|
+
const scaffold = path.join(repoRoot, 'multi-agent-workspace', 'scripts', 'scaffold.js');
|
|
115
|
+
const legacyDecision = path.join(gitFixture, '.harness', 'multi-agent', 'decisions');
|
|
116
|
+
fs.mkdirSync(legacyDecision, { recursive: true });
|
|
117
|
+
fs.writeFileSync(path.join(legacyDecision, 'default-authored.md'), 'migrated from repository-root default\n');
|
|
118
|
+
const res = run(nestedInGit, [scaffold]);
|
|
119
|
+
check(
|
|
120
|
+
'2o. scaffold.js with no --workspace resolves to the repo root from a nested dir',
|
|
121
|
+
res.status === 0 && fs.existsSync(path.join(getRuntimeRoot(gitFixture), 'manifest.json')) &&
|
|
122
|
+
fs.readFileSync(path.join(gitFixture, 'docs', 'adr', 'default-authored.md'), 'utf8') === 'migrated from repository-root default\n',
|
|
123
|
+
`exit=${res.status} stderr=${res.stderr}`
|
|
124
|
+
);
|
|
125
|
+
check(
|
|
126
|
+
'2o. ...and does not scatter runtime files into the repository or nested invocation directory',
|
|
127
|
+
!fs.existsSync(path.join(nestedInGit, '.harness')) && !fs.existsSync(path.join(gitFixture, '.harness')),
|
|
128
|
+
`.harness/ was created inside ${gitFixture} or ${nestedInGit}`
|
|
129
|
+
);
|
|
130
|
+
}
|
|
131
|
+
|
|
132
|
+
{
|
|
133
|
+
const inline = `
|
|
134
|
+
const { getStateRoot, getWorkspaceRoot } = require(${JSON.stringify(path.join(repoRoot, 'hooks', 'scripts', 'lib', 'harness-state.js'))});
|
|
135
|
+
console.log(JSON.stringify({ workspaceRoot: getWorkspaceRoot(), stateRoot: getStateRoot() }));
|
|
136
|
+
`;
|
|
137
|
+
const res = run(nestedInGit, ['-e', inline]);
|
|
138
|
+
let parsed = null;
|
|
139
|
+
try { parsed = JSON.parse(res.stdout); } catch (e) { /* left null */ }
|
|
140
|
+
check(
|
|
141
|
+
'2o. harness-state.js walks up to the repo root from a nested in-repo dir, not the nested dir',
|
|
142
|
+
!!parsed && canonicalPath(parsed.workspaceRoot) === canonicalPath(gitFixture),
|
|
143
|
+
`got ${res.stdout || res.stderr}`
|
|
144
|
+
);
|
|
145
|
+
check(
|
|
146
|
+
'2o. ...and its state dir is under the global home, not under the git fixture at all',
|
|
147
|
+
!!parsed && !canonicalPath(parsed.stateRoot).startsWith(canonicalPath(gitFixture)),
|
|
148
|
+
`stateRoot=${parsed && parsed.stateRoot}`
|
|
149
|
+
);
|
|
150
|
+
}
|
|
151
|
+
|
|
152
|
+
{
|
|
153
|
+
const legacyDir = path.join(gitFixture, '.claude', 'harness-everything', 'state');
|
|
154
|
+
fs.mkdirSync(path.join(legacyDir, 'sessions', 'legacy-session'), { recursive: true });
|
|
155
|
+
fs.writeFileSync(path.join(legacyDir, 'sessions', 'legacy-session', 'handoff-state.json'), '{"status":"idle"}');
|
|
156
|
+
|
|
157
|
+
const inline = `
|
|
158
|
+
const { getStateRoot } = require(${JSON.stringify(path.join(repoRoot, 'hooks', 'scripts', 'lib', 'harness-state.js'))});
|
|
159
|
+
console.log(getStateRoot());
|
|
160
|
+
`;
|
|
161
|
+
const res = run(gitFixture, ['-e', inline], { ...baseEnv, CLAUDE: '1' });
|
|
162
|
+
const newStateRoot = (res.stdout || '').trim();
|
|
163
|
+
|
|
164
|
+
check(
|
|
165
|
+
'2o. pre-existing scattered state migrates into the new global location',
|
|
166
|
+
newStateRoot && fs.existsSync(path.join(newStateRoot, 'sessions', 'legacy-session', 'handoff-state.json')) && fs.existsSync(path.join(newStateRoot, '.migrated-from')),
|
|
167
|
+
`exit=${res.status} stderr=${res.stderr} newStateRoot=${newStateRoot}`
|
|
168
|
+
);
|
|
169
|
+
check(
|
|
170
|
+
'2o. ...and the old in-repo location is gone afterward',
|
|
171
|
+
!fs.existsSync(legacyDir),
|
|
172
|
+
`${legacyDir} still exists`
|
|
173
|
+
);
|
|
174
|
+
}
|
|
175
|
+
|
|
176
|
+
fs.rmSync(gitFixture, { recursive: true, force: true });
|
|
177
|
+
fs.rmSync(fakeHome, { recursive: true, force: true });
|
|
178
|
+
|
|
179
|
+
if (failures > 0) {
|
|
180
|
+
console.error(`\n${failures} check(s) failed.`);
|
|
181
|
+
process.exit(1);
|
|
182
|
+
}
|
|
183
|
+
process.exit(0);
|