# RDC Skill Test Suite Matrix

Current coverage: 30 manifests for 30 skill directories.

The manifest layer verifies each skill can be started from a realistic caller prompt in an isolated `RDC_TEST=1` sandbox. The acceptance harness can run either Claude (`--engine claude`) or Codex (`--engine codex`) against the same manifests and records the engine stream, extracted tool calls, stdout/stderr artifacts, rendered assistant output, failures, lessons learned, and next build optimizations under `.rdc/reports/`.

`rdc:channel-formatter` currently has the strongest content acceptance fixture: it asserts source-grounded social-pack output, forbidden-claim absence, and observable corpus or web-search tool routing. `rdc:help` asserts the public MCP/curl caller surface and structured `format:"json"` discovery path. `rdc:self-test` validates the test-tier/evidence language. Planning/reporting skills (`rdc:preplan`, `rdc:plan`, `rdc:handoff`, `rdc:prototype`, `rdc:review`, `rdc:report`) assert their required artifacts and safety boundaries. Operational/content skills (`rdc:brochure`, `rdc:convert`, `lifeai-brochure-author`, `rdc:build`, `rdc:fixit`, `rdc:fs-mcp`, `rdc:workitems`) assert artifact paths, dispatch gates, scoped commits, filesystem boundaries, and sandboxed work-item behavior. Coordination/design/maintenance skills (`rdc:co-develop`, `rdc:collab`, `rdc:design`, `rdc:extract-verifier-rules`, `rdc:housekeeping`, `rdc:overnight`, `rdc:rpms-filemap`) assert relay safety, token/reference use, verifier corpus safety, maintenance/reporting, unattended sentinel behavior, and canonical file-map rules. All manifests now include a deeper acceptance block beyond the basic executable smoke fixture.

| Skill | Manifest | Fixture prompt class | Assertions | Acceptance depth |
|---|---|---|---|---|
| `rdc:brochure` | `rdc-brochure.test.json` | HTML-to-PDF brochure fixture | `exit_code`, `stdout_contains` | PDF artifact, page/size/source output, and no upload/deploy/source-modification checks |
| `rdc:build` | `rdc-build.test.json` | Unattended build from sandbox label | `commits_made`, `exit_code`, `stdout_contains` | Dispatch guide, code-review, validator, sandbox, and no-push/no-skip checks |
| `rdc:channel-formatter` | `rdc-channel-formatter.test.json` | Long article to social content pack | `exit_code`, `stdout_contains` | Output contains, output not contains, tool-call routing |
| `rdc:co-develop` | `rdc-co-develop.test.json` | Coordination status | `commits_made`, `exit_code`, `stdout_contains` | codevelop status, sandbox skip, no chitchat/live-send negative checks |
| `rdc:collab` | `rdc-collab.test.json` | Claude session relay fixture | `commits_made`, `exit_code`, `stdout_contains` | chitchat/SSE session handling, sandbox skip, and no-reply/no-push negative checks |
| `rdc:convert` | `rdc-convert.test.json` | Markdown-to-Word conversion fixture | `exit_code`, `stdout_contains` | build-corpus command surface, Word/Markdown flags, and no-GUI/global-install/commit checks |
| `rdc:deploy` | `rdc-deploy.test.json` | Deployment diagnosis | `commits_made`, `exit_code`, `stdout_contains` | Read-only diagnose output and destructive deploy/DNS negative checks |
| `rdc:edit` | `rdc-edit.test.json` | Local editor launch | `exit_code`, `stdout_contains` | Resolved target, editor host URL, and no-unknown-target negative checks |
| `rdc:design` | `rdc-design.test.json` | Studio palette audit | `exit_code`, `stdout_contains` | Studio/token/Rampa references and external-write negative checks |
| `rdc:fixit` | `rdc-fixit.test.json` | Tiny sandbox typo fix | `commits_made`, `exit_code`, `stdout_contains` | Scope, code-review, specific-file staging, and broad-add negative checks |
| `rdc:fs-mcp` | `rdc-fs-mcp.test.json` | File-system bridge read fixture | `commits_made`, `exit_code`, `stdout_contains` | Local-vs-MCP access boundary, live FS, no-preview/no-overwrite checks, and read-tool routing |
| `rdc:handoff` | `rdc-handoff.test.json` | Stub work handoff | `exit_code`, `stdout_contains` | Plan path, DoD, work-item handoff, and placeholder negative checks |
| `rdc:help` | `rdc-help.test.json` | Help menu rendering | `commits_made`, `exit_code`, `stdout_contains` | MCP/curl output, structured JSON fetch, dev-endpoint negative checks |
| `rdc:housekeeping` | `rdc-housekeeping.test.json` | Read-only housekeeping audit | `commits_made`, `exit_code`, `stdout_contains` | PUBLISH/CLAUDE/orphan/report checks and no-fix/no-Coolify-write negative checks |
| `lifeai-brochure-author` | `rdc-lifeai-brochure-author.test.json` | JSX compliance review fixture | `exit_code`, `stdout_contains` | brochure-kit component guidance and raw HTML/style/token negative checks |
| `rdc:overnight` | `rdc-overnight.test.json` | Label-based overnight queue drain | `exit_code`, `stdout_contains` | overnight sentinel, clauth preflight, full chain, and no-wide-add/no-lock-left negative checks |
| `rdc:plan` | `rdc-plan.test.json` | Health-endpoint planning prompt | `exit_code`, `stdout_contains` | Design-decision, decomposition-matrix, work-package, and coarse-check negative checks |
| `rdc:preplan` | `rdc-preplan.test.json` | Rate-limiter research prompt | `commits_made`, `exit_code`, `stdout_contains` | Research artifact, comparison, unknowns, recommendation, and research-tool routing checks |
| `rdc:prototype` | `rdc-prototype.test.json` | Tiny component prototype prompt | `exit_code`, `stdout_contains` | Local TSX prototype artifact, registry, test-mode, and production-write negative checks |
| `rdc:brochurify` | `rdc-rdc-brochurify.test.json` | Read-only markdown Brochurify fixture | `exit_code`, `stdout_contains` | Sandbox output and no-follow-up negative checks |
| `rdc:extract-verifier-rules` | `rdc-rdc-extract-verifier-rules.test.json` | Enhancement-log verifier fixture | `exit_code`, `stdout_contains` | Candidate, false-positive, known-good, recall, and no-auto-merge/PR negative checks |
| `rdc:release` | `rdc-release.test.json` | Dry-run package release | `commits_made`, `exit_code`, `stdout_contains` | Dry-run release checklist and force/bypass negative checks |
| `rdc:report` | `rdc-report.test.json` | Unattended report generation | `exit_code`, `stdout_contains` | Report status block, counts, next recommendation, and dump/push negative checks |
| `rdc:review` | `rdc-review.test.json` | Unattended review gate | `exit_code`, `stdout_contains` | Code-review, verify, fresh-evidence, tsc, and forbidden-build negative checks |
| `rdc:rpms-filemap` | `rdc-rpms-filemap.test.json` | Canonical location lookup | `commits_made`, `exit_code`, `stdout_contains` | Rule #1, context door, one-home rules, and no-hand-edit/second-home negative checks |
| `rdc:self-test` | `rdc-self-test.test.json` | Strict self-test prompt | `commits_made`, `exit_code`, `stdout_contains` | Strict validation evidence, hook/guide checks, and no-relaxed-linter negative check |
| `rdc:status` | `rdc-status.test.json` | Read-only status snapshot | `commits_made`, `exit_code`, `stdout_contains` | Read-only status output and raw-MCP/write negative checks |
| `rdc:terminal-config` | `rdc-terminal-config.test.json` | Hidden-window launch policy audit | `commits_made`, `exit_code`, `stdout_contains` | Hidden-window output and focus/collapse API negative checks |
| `rdc:watch` | `rdc-watch.test.json` | Watcher initialization prompt | `exit_code`, `stdout_contains` | Test-mode no-open output, tool-call routing, and focus-event negative checks |
| `rdc:workitems` | `rdc-workitems.test.json` | Work-item epic listing | `commits_made`, `exit_code`, `stdout_contains` | Read-only epics query, sandbox, and work-item write negative checks |

## What The Tests Accomplish

- Verify every installed RDC skill has at least one executable caller-facing prompt.
- Exercise the `RDC_TEST=1` sandbox contract so build tests do not publish, deploy, or mutate production services.
- Capture observed engine events and extracted tool calls for periodic inspection.
- Preserve stdout, stderr, and rendered assistant output as report artifacts.
- Make content-producing skills eligible for deeper source-fidelity checks through the optional `acceptance` block.

## Next Coverage Upgrades

- Run full scheduled Claude and Codex sweeps regularly, not only changed-skill acceptance, so transcripts and lessons stay fresh.
- Add richer artifact-content checks for output-producing skills (`rdc:brochure`, `rdc:convert`, `rdc:prototype`, `rdc:report`).
- Promote recurring transcript review findings into manifest assertions instead of relying on human inspection.