Skip to content

Skills Catalog ​

Skills are on-demand procedural bundles containing contracts, executable scripts, verification fixtures, and scaffolding templates.

Installed via APM as directory bundles under .agents/skills/<name>/ (Copilot and Cursor) and .claude/skills/<name>/ (Claude Code).

Overview Table ​

Skill DirectoryTypePaired PrimitiveDescription
skills/think/Protocol & Validatorthink instruction, Thinker agentContract and deterministic validator (validate-think.mjs) for the Think record (think.md).
skills/plan/Protocol & Validatorplan instruction, Planner agentContract and deterministic validator (validate-plan.mjs) for the Plan record (plan.md), including cross-checks against think.md.
skills/code/Protocol & ValidatorCode Agent, code instructionAgent branch guard (branch-guard.mjs) and pull request contract (validate-pr.mjs): links, acceptance-criteria traceability, dependencies, blast radius, and the human-validation record by tier.
skills/prove/Protocol, Validator & ControlsProve Validator, Gate Fixer, prove instructionProve record contract (specs/issue-<n>/prove.md), validator (validate-prove.mjs), and all 8 deterministic control scripts (control-guard, secret-guard, dependency-guard, scope-guard, task-gate, criteria-verify, test-probe, preview-smoke).
skills/release-plan/Protocol & Validatorrelease instruction, Release PlannerPer-change release.md contract and validator, plus the deterministic rollout decision (promote, hold, rollback).
skills/hypothesis-ledger/Protocol & ValidatorLearn AnalystHypothesis ledger with deterministic verdict proposals and feature-deletion tracking.
skills/incident-record/Protocol & ValidatorLearn Analyst, learn instructionIncident record contract and validator: mitigation before diagnosis, failing test first, review within five working days.
skills/telemetry-scan/Behavioral ProbeLearn AnalystPredictive monitoring from exported telemetry: anomalies, trends, breaches, and the deploys that precede them.
skills/learn-signals/Protocol & EmitterLearn Analyst, triage skillTurns incidents, invalidated hypotheses, and findings into Triage issue bodies.
skills/context-health/Guard & AuditLearn AnalystPersistent Context lint and a corrections log that proposes new artifacts after two corrections.
skills/decision-record/Protocol & Validatordecision-record instruction, Decision ScribeAttributed autonomous decisions and escalations, with genuine human signatures supported when requested.
skills/flow-metrics/MetricsLearn Analyst, retrospectivesThe factory's metrics computed from a JSONL event log.
skills/meeting-pack/Protocol & Validatormeeting-pack instruction, Meeting Analyst agentMeeting Pack contract, plain-text transcript normalizer, deterministic validator (verbatim quotes, glossary, confirmations), and Triage emission.
skills/triage/Protocol & Validatortriage instruction, Triage Analyst agentTriage contract, deterministic validator (evidence, scorecard, decision rules, ledger), and backlog ranking for triage.md artifacts.
skills/spec-from-code/Protocol & Validatorspec-from-codeTechnical specification contract under docs/specs/ with automated schema validation.
skills/vitepress-docs/Standards & ScaffoldingVitePress Docs agentVitePress site rules and templates: package.json, config.mjs, dev Dockerfile, and Pages workflow.
skills/playwright-demo-videos/Standards & ScaffoldingDemo Videos agentDemo recording rules and templates: Playwright config, Director (fake cursor, gestures), scenarios, ffmpeg encoder, VitePress video section.
skills/node-express-examples/Scaffolding Packnode-express instructionExpress scaffolding: route handlers, controllers, middlewares, and error envelopes.
skills/postgres-liquibase-examples/Scaffolding Packpostgres-liquibase instructionLiquibase XML/YAML changelogs, rollback blocks, audit triggers, and soft-delete SQL.
skills/docker-examples/Scaffolding Packdocker instructionMulti-stage Dockerfiles, Traefik dynamic labels, BuildKit secret mounts, and Compose files.
skills/angular-admin-examples/Scaffolding Packangular-admin instructionAngular entity slices: CRUD tables, PrimeNG dialogs, reactive form services, and ACL guards.
skills/k6-performance-tests/Standards & ScaffoldingPerformance Tester agentModular k6 test architecture, thresholds, authentication, stages, and Docker runners.
skills/restler-fuzzing-tests/Standards & ScaffoldingFuzz Tester agentRESTler OpenAPI compiler rules, fuzzing modes, token refresh, and headless CI runner scripts.
skills/contract-validator/Schema Contractspec-from-codeDeclarative schema validator for Markdown and structured artifacts without custom scripts.
skills/rollback-probe/Behavioral Probepostgres-liquibase, database migrationsCertifies migration reversibility (Forward $\rightarrow$ Rollback $\rightarrow$ Forward) and idempotency.
skills/migration-docs/Protocol & Validatormigration-docs instruction, Migration Documenter agentContract and deterministic validator for the infrastructure migration records (assessment.md, migration-plan.md, runbook.md) and their cross-checks.

Skill Deep Dives ​

1. Playwright Demo Videos (skills/playwright-demo-videos/) ​

  • Category: Standards & Scaffolding
  • Contents: SKILL.md (filming, reset, encoding, and embedding rules) and references/examples.md (demo playwright.config.ts, support/director.ts with fake cursor and human-speed gestures, login and editing scenario templates with off-camera reset, to-mp4.mjs ffmpeg encoder, VitePress DemoVideos.vue home section).
  • Workflow: npm run demos in the app, i.e. playwright test --config=demos/playwright.config.ts && node demos/to-mp4.mjs <output-dir>.

2. Spec from Code (skills/spec-from-code/) ​

  • Category: Protocol & Validator
  • Contents:
    • SKILL.md: Step-by-step extraction workflow.
    • references/spec-contract.md: Spec file schema for area specs, component sections (Purpose, Inputs, Outputs, Side effects, Tests), and Ambiguous or undocumented sections.
    • scripts/spec-docs.mjs: Automated validator asserting that all spec files conform to file shape, heading rules, component fields, and max line limits.
    • scripts/spec-docs.test.mjs: Node.js test suite asserting spec validation coverage.
  • Workflow:
    bash
    # Validate technical specifications in CI
    node .agents/skills/spec-from-code/scripts/spec-docs.mjs --dir docs/specs

3. VitePress Docs (skills/vitepress-docs/) ​

  • Category: Standards & Scaffolding
  • Contents:
    • SKILL.md: Normative VitePress layout, stack, branding, sidebar rules, and path resolution.
    • references/examples.md: Complete copy-ready templates for package.json, config.mjs, index.md, dev dockerfile, and GitHub Pages workflow.
  • Usage: Scaffolding agents substitute <docs-root> and copy templates directly to set up documentation sites like this one.

4. Node.js Express Examples (skills/node-express-examples/) ​

  • Category: Scaffolding Pack
  • Contents:
    • Scaffolding templates for route handlers, Joi/Zod request validation middlewares, centralized error handling, and Supertest integration tests.

5. PostgreSQL & Liquibase Examples (skills/postgres-liquibase-examples/) ​

  • Category: Scaffolding Pack
  • Contents:
    • Templates for audited table creations, auto-updating updated_at triggers, soft-delete rules, and corresponding rollback tags.

6. Docker Examples (skills/docker-examples/) ​

  • Category: Scaffolding Pack
  • Contents:
    • Alpine Node, Go, and Python Dockerfile snippets using BuildKit secret caching and Traefik Compose declarations.

7. Angular Admin Examples (skills/angular-admin-examples/) ​

  • Category: Scaffolding Pack
  • Contents:
    • Feature-sliced entity templates, PrimeNG Table configs, column sorting, pagination, and ACL-aware button directives.

8. k6 Performance Tests (skills/k6-performance-tests/) ​

  • Category: Standards & Scaffolding
  • Contents:
    • SKILL.md & references/implementation-patterns.md: Normative architecture, threshold rules, authentication lifecycle, and templates for load, soak, and stress testing.

9. RESTler Fuzzing Tests (skills/restler-fuzzing-tests/) ​

  • Category: Standards & Scaffolding
  • Contents:
    • SKILL.md & references/implementation-patterns.md: Normative rules and templates for OpenAPI compiler configurations, token-refresh authentication, data safety, and headless CI runner scripts.

10. Think (skills/think/) & Plan (skills/plan/) ​

  • Category: Protocol & Validator
  • Contents:
    • skills/think/SKILL.md & references/think-contract.md: Authoritative schema for think.md (options, chosen approach, system flow, constraints, blast radius, decision record, Persistent Context updates).
    • skills/think/scripts/validate-think.mjs: Standalone deterministic Node.js validator for think.md (--think or --dir, with --stage draft|approved).
    • skills/plan/SKILL.md & references/plan-contract.md: Authoritative schema for plan.md (file scope, acceptance criteria, steps with Files, Action, Verification, Bootable Check, Serves, test plan, blast radius check).
    • skills/plan/scripts/validate-plan.mjs: Standalone deterministic Node.js validator for plan.md (--plan or --dir, with --think cross-check and --stage draft|approved).
    • scripts/*.test.mjs and scripts/*-fixtures.mjs: Test suites and valid fixtures for each skill.
  • Workflow:
    bash
    # Validate think.md in CI or local harness
    node .agents/skills/think/scripts/validate-think.mjs --think specs/issue-42/think.md
    
    # Validate plan.md and cross-check against think.md
    node .agents/skills/plan/scripts/validate-plan.mjs --dir specs/issue-42

11. Prove (skills/prove/) ​

  • Category: Protocol, Validator & Controls
  • Contents:
    • SKILL.md & references/prove-contract.md: Schema for specs/issue-<n>/prove.md (header links, 8 required deterministic controls, criteria verification table cross-checked against plan.md, semantic review across Client Intent, Behavioral Experience, and Safety & Blast Radius, numbered findings, and PASS|FAIL verdict) and scripts/validate-prove.mjs (--stage draft|pass).
    • scripts/control-guard.mjs & references/control-guard-contract.md: Control integrity guard (rules R1-R8: deleted tests, skip markers, lowered thresholds, weakened workflows, protected paths, suppressions, production credentials in agent workflows, removed assertions).
    • scripts/secret-guard.mjs & references/secret-contract.md: Fast regex and entropy scanner asserting diffs or source files contain zero hardcoded secrets.
    • scripts/dependency-guard.mjs & references/dependency-contract.md: Manifest invariant guard inspecting package.json diffs to block unvetted or banned libraries.
    • scripts/scope-guard.mjs & references/guard-contract.md: Blast-radius and churn guard asserting diffs stay within permitted --scope globs, forbidden file rules, and churn limits.
    • scripts/task-gate.mjs & references/gate-contract.md: Multi-stage verification runner executing sequential checks with fail-fast semantics and decisive failure snippet extraction.
    • scripts/verify-criteria.mjs & references/criteria-verify-contract.md: Criteria Test Verification gate mapping tests to AC-<n> from JUnit XML, Jest/Vitest JSON, or TAP / runner logs.
    • scripts/test-probe.mjs & references/probe-contract.md: Anti-tautology probe temporarily unpatching source code to prove tests genuinely fail without the fix.
    • scripts/preview-smoke.mjs & references/preview-smoke-contract.md: Preview smoke probe checking routes (status, JSON path, latency) of a running preview.
  • Workflow:
    bash
    # Run Prove controls and validate prove.md
    node .agents/skills/prove/scripts/scope-guard.mjs --git --base origin/main --scope "src/auth/**"
    node .agents/skills/prove/scripts/secret-guard.mjs --git --base origin/main
    node .agents/skills/prove/scripts/dependency-guard.mjs --git --base origin/main --allow "zod,dotenv"
    node .agents/skills/prove/scripts/control-guard.mjs --git origin/main
    node .agents/skills/prove/scripts/task-gate.mjs --stage "lint:npm run lint" --stage "tests:npm test" --json
    node .agents/skills/prove/scripts/verify-criteria.mjs --plan specs/issue-42/plan.md --report junit.xml
    node .agents/skills/prove/scripts/test-probe.mjs --test-cmd "npm test -- tests/auth.test.js" --source "src/auth/service.js" --base origin/main
    node .agents/skills/prove/scripts/preview-smoke.mjs --url http://localhost:3000
    node .agents/skills/prove/scripts/validate-prove.mjs --dir specs/issue-42 --stage pass

12. Contract Validator (skills/contract-validator/) ​

  • Category: Schema Contract
  • Contents:
    • SKILL.md: Declarative document validation workflow.
    • references/contract-schema.md: Specification for *.contract.json rules (markers, headings, sequence, max lines).
    • scripts/validate-contract.mjs: Generic validator evaluating Markdown/JSON files against declarative contracts without writing custom scripts.
    • scripts/validate-contract.test.mjs: Test suite covering marker detection, heading sequences, line overflow, and forbidden patterns.
  • Workflow:
    bash
    # Declaratively validate an artifact against a contract
    node .agents/skills/contract-validator/scripts/validate-contract.mjs --contract contracts/report.contract.json --file docs/report.md

13. Rollback Probe (skills/rollback-probe/) ​

  • Category: Behavioral Probe
  • Contents:
    • SKILL.md: Migration reversibility and idempotency verification workflow.
    • references/rollback-contract.md: Roundtrip lifecycle specification (Forward $\rightarrow$ Rollback $\rightarrow$ Forward Re-apply).
    • scripts/rollback-probe.mjs: CLI automating migration application, rollback, and re-apply to guarantee clean inverse state.
    • scripts/rollback-probe.test.mjs: Test suite covering forward failure, rollback failure, and idempotency verification.
  • Workflow:
    bash
    # Certify database migration reversibility and idempotency
    node .agents/skills/rollback-probe/scripts/rollback-probe.mjs \
      --forward "npm run migrate:up" \
      --rollback "npm run migrate:down"

14. Triage (skills/triage/) ​

  • Category: Protocol & Validator
  • Contents:
    • SKILL.md: Step-by-step guidance on authoring, validating, and ranking triage.md artifacts.
    • references/triage-contract.md: Authoritative schema for triage.md (backlog issue, evidence, system mapping, scorecard, decision), the validation stages, the triage ledger, and the strategic weights.
    • scripts/validate-triage.mjs: Deterministic Node.js validator for the proposed and final stages, with verbatim quote verification (--source) and ledger checks (--ledger).
    • scripts/rank-backlog.mjs: Deterministic ranking of scored triage issues for the weekly review, using Product Owner weights.
    • scripts/*.test.mjs and scripts/triage-fixture.mjs: Test suites and the shared valid fixture.
  • Workflow:
    bash
    # Gate the AI output, then the Product Owner decision
    node .agents/skills/triage/scripts/validate-triage.mjs --dir specs/issue-42 --stage proposed
    node .agents/skills/triage/scripts/validate-triage.mjs --dir specs/issue-42 --stage final
    # Rank the backlog for the weekly Triage review
    node .agents/skills/triage/scripts/rank-backlog.mjs --dir specs --weights specs/triage/weights.json

15. Meeting Pack (skills/meeting-pack/) ​

  • Category: Protocol & Validator
  • Contents:
    • SKILL.md: Step-by-step guidance on normalizing, building, validating, confirming, and handing off a Meeting Pack.
    • references/meeting-pack-contract.md: Authoritative schema for transcript.md, pack.md, the validation stages, the glossary, and the Triage handoff.
    • scripts/normalize-transcript.mjs: Converts a plain-text transcript export and a speaker roster into the immutable transcript.md and prints its SHA-256.
    • scripts/validate-meeting.mjs: Deterministic validator for the transcript, the pack (draft and confirmed stages), and docs/glossary.md, including verbatim quote verification and the transcript digest.
    • scripts/emit-candidates.mjs: Turns a confirmed pack into backlog issue bodies, candidates.json, and glossary-additions.md.
    • scripts/*.test.mjs and scripts/meeting-fixtures.mjs: Test suites and the shared valid fixtures.
  • Workflow:
    bash
    node .agents/skills/meeting-pack/scripts/validate-meeting.mjs --dir meetings/2026-09-24-acme-discovery --glossary docs/glossary.md
    node .agents/skills/meeting-pack/scripts/validate-meeting.mjs --dir meetings/2026-09-24-acme-discovery --glossary docs/glossary.md --stage confirmed
    node .agents/skills/meeting-pack/scripts/emit-candidates.mjs --dir meetings/2026-09-24-acme-discovery --out build/candidates

16. Code (skills/code/) ​

  • Category: Protocol & Validator
  • Contents: Agent branch guard (scripts/branch-guard.mjs), PR contract (references/code-contract.md), template (references/pull_request_template.md), and PR validator (scripts/validate-pr.mjs). Code checks short-lived branch naming/age, scope, traceability, and dependencies. Prove supports all-tier autonomous assessment with --allow-auto-prove, retaining High/Critical rollback evidence and Critical rollout references. Genuine human formats remain supported.
  • Workflow: node .agents/skills/code/scripts/branch-guard.mjs --git --base main --max-age-days 5 and node .agents/skills/code/scripts/validate-pr.mjs --body pr.md --dir specs/issue-42 --stage prove

17. Release Plan (skills/release-plan/) ​

  • Category: Protocol & Validator
  • Contents: references/release-contract.md, references/release.yml (tag-triggered workflow template), scripts/validate-release.mjs, and scripts/evaluate-rollout.mjs (exit 0 promote, 2 hold, 1 rollback).
  • Workflow: node .agents/skills/release-plan/scripts/validate-release.mjs --dir specs/issue-42

18. Hypothesis Ledger (skills/hypothesis-ledger/) ​

  • Category: Protocol & Validator
  • Contents: scripts/ledger.mjs (validate, add, due, propose, stats). propose computes a verdict from baseline, target, and observed value and never writes it.
  • Workflow: node .agents/skills/hypothesis-ledger/scripts/ledger.mjs due --file specs/learn/hypotheses.md

19. Incident Record (skills/incident-record/) ​

  • Category: Protocol & Validator
  • Contents: scripts/validate-incident.mjs enforces timestamp order, mitigation before diagnosis, a reproducing test, the five-working-day review, and control hardening, and reports time to restore.
  • Workflow: node .agents/skills/incident-record/scripts/validate-incident.mjs --file docs/incidents/2026-10-01-checkout-errors.md

20. Telemetry Scan (skills/telemetry-scan/) ​

  • Category: Behavioral Probe
  • Contents: scripts/telemetry-scan.mjs finds anomalies, trends toward a threshold, and breaches in exported series, and lists the deploys that precede them. Exit 3 means a high-severity finding.
  • Workflow: node .agents/skills/telemetry-scan/scripts/telemetry-scan.mjs --series series.json --thresholds thresholds.json

21. Learn Signals (skills/learn-signals/) ​

  • Category: Protocol & Emitter
  • Contents: scripts/emit-learn-issues.mjs writes Triage issue bodies (needs:triage, source:learn, learn-signal marker) from validated incidents, ledger verdicts, and telemetry findings.
  • Workflow: node .agents/skills/learn-signals/scripts/emit-learn-issues.mjs --out build/signals --incident docs/incidents/2026-10-01-checkout-errors.md

22. Context Health (skills/context-health/) ​

  • Category: Guard & Audit
  • Contents: scripts/context-lint.mjs (missing description, broad applyTo, oversized or near-duplicate instructions, missing references, secrets, staleness, growth) and scripts/context-corrections.mjs (a log that proposes an artifact once a topic is corrected or re-explained twice).
  • Workflow: node .agents/skills/context-health/scripts/context-lint.mjs --root .

23. Decision Record (skills/decision-record/) ​

  • Category: Protocol & Validator
  • Contents: references/decision-record-contract.md, scripts/validate-decision.mjs (records, numbering, Think cross-check), and scripts/escalations.mjs (open escalations whose default date has passed).
  • Workflow: node .agents/skills/decision-record/scripts/validate-decision.mjs --dir docs/decisions

24. Flow Metrics (skills/flow-metrics/) ​

  • Category: Metrics
  • Contents: references/metrics-contract.md, scripts/emit-event.mjs, and scripts/flow-metrics.mjs, which reports every factory metric against its target from metrics/events.jsonl. Lines of code, story points, and editor suggestion acceptance are never computed.
  • Workflow: node .agents/skills/flow-metrics/scripts/flow-metrics.mjs --events metrics/events.jsonl

25. Migration Docs (skills/migration-docs/) ​

  • Category: Protocol & Validator
  • Contents: SKILL.md, references/migration-contract.md (record structure, Application Change Policy, cross-checks), and scripts/validate-migration.mjs with its fixtures and tests.
  • Workflow: node .agents/skills/migration-docs/scripts/validate-migration.mjs --dir docs/migration/<name>.

Published and maintained by ALTEN