Skip to content

cskwork/supergoal-skill / skills/supergoal

supergoal

Unverified

Route objective-driven development, verification, planning, "teach/explain", codebase learning, and skill evaluation through the Supergoal workflow. Use when Supergoal is requested or an autonomous delivery workflow is needed.

FileKindSize
SKILL.mdskill md12.8 KB
agents/analyst.mdmarkdown1.5 KB
agents/architect.mdmarkdown1.7 KB
agents/code-reviewer.mdmarkdown2.5 KB
agents/db-reader.mdmarkdown2.0 KB
agents/debugger.mdmarkdown2.0 KB
agents/designer.mdmarkdown6.3 KB
agents/executor.mdmarkdown2.8 KB
agents/explore.mdmarkdown2.0 KB
agents/qa-auditor.mdmarkdown4.3 KB
agents/qa-tester.mdmarkdown3.2 KB
agents/security-reviewer.mdmarkdown1.2 KB
agents/skill-forger.mdmarkdown1.6 KB
agents/skill-miner.mdmarkdown1.6 KB
reference/agent-browser.mdmarkdown2.8 KB
reference/arch.mdmarkdown5.1 KB
reference/archify.mdmarkdown3.6 KB
reference/db-access.mdmarkdown2.9 KB
reference/debugging.mdmarkdown6.5 KB
reference/delivery-gate.mdmarkdown6.6 KB
reference/domain-context.mdmarkdown7.8 KB
reference/domain-rules.mdmarkdown1.6 KB
reference/engagement.mdmarkdown3.8 KB
reference/functional-ui.mdmarkdown6.4 KB
reference/harness-eval.mdmarkdown28.7 KB
reference/interview.mdmarkdown7.7 KB
reference/learn-domain.mdmarkdown12.7 KB
reference/market-research.mdmarkdown1.7 KB
reference/observability.mdmarkdown5.0 KB
reference/plan-grounding.mdmarkdown4.3 KB
reference/playwright-cli.mdmarkdown2.6 KB
reference/prototype.mdmarkdown4.5 KB
reference/qa-only.mdmarkdown12.7 KB
reference/qa.mdmarkdown7.9 KB
reference/reporting.mdmarkdown1.5 KB
reference/research.mdmarkdown2.7 KB
reference/review-only.mdmarkdown2.4 KB
reference/role-loop.mdmarkdown26.7 KB
reference/rules.mdmarkdown982 B
reference/skill-mine.mdmarkdown7.6 KB
reference/taste-aesthetics.mdmarkdown7.5 KB
reference/taste-skill-v2.mdmarkdown18.8 KB
reference/taste-sources.mdmarkdown3.0 KB
reference/teach.mdmarkdown28.0 KB
reference/ui-ux.mdmarkdown4.0 KB
reference/vercel-host.mdmarkdown3.6 KB
reference/wayfinder.mdmarkdown7.2 KB
templates/GOAL.mdmarkdown1.3 KB
templates/PLAN.mdmarkdown1.4 KB
templates/QA.mdmarkdown1.2 KB
templates/R-LOOP.mdmarkdown569 B
templates/Z-DONE.mdmarkdown301 B
templates/arch-report.htmlother8.1 KB
templates/commit-gate.shexecutablescript11.6 KB
templates/contrast-gate.mjsexecutablescript4.3 KB
templates/db-access/.env.exampleother451 B
templates/db-access/check-connection.shexecutablescript134 B
templates/db-access/db-access.cmdscript39 B
templates/db-access/db-access.mjsexecutablescript5.2 KB
templates/db-access/read-only-query.shexecutablescript138 B
templates/db-access/schema-summary.shexecutablescript132 B
templates/domain-agent/code-map.mdmarkdown744 B
templates/domain-agent/config.jsonconfig291 B
templates/domain-agent/decisions/README.mdmarkdown270 B
templates/domain-agent/flows/README.mdmarkdown800 B
templates/domain-agent/freshness.mdmarkdown1.2 KB
templates/domain-agent/glossary.mdmarkdown781 B
templates/domain-agent/index.mdmarkdown1.3 KB
templates/domain-agent/invariants.mdmarkdown504 B
templates/domain-agent/qa/README.mdmarkdown879 B
templates/domain-agent/qa/nav-map.mdmarkdown1.2 KB
templates/domain-agent/test-map.mdmarkdown432 B
templates/domain-agent/tickets/README.mdmarkdown204 B
templates/domain-onboarding.htmlother9.2 KB
templates/harness-eval-case.yamlconfig4.4 KB
templates/harness-eval-cases/authored/authored-underspec-001-deepmerge.yamlconfig2.2 KB
templates/harness-eval-cases/authored/authored-underspec-002-csvline.yamlconfig2.0 KB
templates/harness-eval-cases/authored/authored-underspec-003-authz-cache.yamlconfig1.6 KB
templates/harness-eval-cases/fixtures/README.mdmarkdown2.8 KB
templates/harness-eval-cases/fixtures/revfactory-case-002-async-race/package.jsonmanifest124 B
templates/harness-eval-cases/fixtures/revfactory-case-002-async-race/src/cache.mjsscript494 B
templates/harness-eval-cases/fixtures/revfactory-case-002-async-race/test/cache.hidden.test.mjsscript2.2 KB
templates/harness-eval-cases/fixtures/revfactory-case-002-async-race/test/cache.visible.test.mjsscript950 B
templates/harness-eval-cases/fixtures/revfactory-case-003-refactoring/package.jsonmanifest125 B
templates/harness-eval-cases/fixtures/revfactory-case-003-refactoring/src/order.mjsscript1.3 KB
templates/harness-eval-cases/fixtures/revfactory-case-003-refactoring/test/order.hidden.test.mjsscript2.2 KB
templates/harness-eval-cases/fixtures/revfactory-case-003-refactoring/test/order.visible.test.mjsscript1.2 KB
templates/harness-eval-cases/fixtures/underspec-001-deepmerge/package.jsonmanifest124 B
templates/harness-eval-cases/fixtures/underspec-001-deepmerge/src/merge.mjsscript84 B
templates/harness-eval-cases/fixtures/underspec-001-deepmerge/test/merge.hidden.test.mjsscript1013 B
templates/harness-eval-cases/fixtures/underspec-001-deepmerge/test/merge.visible.test.mjsscript276 B
templates/harness-eval-cases/fixtures/underspec-002-csvline/package.jsonmanifest122 B
templates/harness-eval-cases/fixtures/underspec-002-csvline/src/csv.mjsscript77 B
templates/harness-eval-cases/fixtures/underspec-002-csvline/test/csv.hidden.test.mjsscript804 B
templates/harness-eval-cases/fixtures/underspec-002-csvline/test/csv.visible.test.mjsscript240 B
templates/harness-eval-cases/fixtures/underspec-003-authz-cache/package.jsonmanifest126 B
templates/harness-eval-cases/fixtures/underspec-003-authz-cache/src/authorizer.mjsscript601 B
templates/harness-eval-cases/fixtures/underspec-003-authz-cache/test/authorizer.hidden.test.mjsscript3.8 KB
templates/harness-eval-cases/fixtures/underspec-003-authz-cache/test/authorizer.visible.test.mjsscript1.2 KB
templates/harness-eval-cases/revfactory-case-001-rest-api.yamlconfig2.1 KB
templates/harness-eval-cases/revfactory-case-002-bug-fix.yamlconfig2.4 KB
templates/harness-eval-cases/revfactory-case-003-refactoring.yamlconfig2.4 KB
templates/harness-eval-cases/revfactory-case-004-documentation.yamlconfig2.1 KB
templates/harness-eval-cases/revfactory-case-005-complex.yamlconfig2.0 KB
templates/harness-eval-cases/revfactory-case-006-research.yamlconfig2.1 KB
templates/harness-eval-cases/revfactory-case-007-interpreter.yamlconfig2.1 KB
templates/harness-eval-cases/revfactory-case-008-microservice.yamlconfig2.1 KB
templates/harness-eval-cases/revfactory-case-009-sql-engine.yamlconfig2.0 KB
templates/harness-eval-cases/revfactory-case-010-crdt.yamlconfig2.0 KB
templates/harness-eval-cases/revfactory-case-011-raft-kv.yamlconfig2.1 KB
templates/harness-eval-cases/revfactory-case-012-spreadsheet.yamlconfig2.1 KB
templates/harness-eval-cases/revfactory-case-013-bytecode-vm.yamlconfig2.1 KB
templates/harness-eval-cases/revfactory-case-014-event-sourcing.yamlconfig2.1 KB
templates/harness-eval-cases/revfactory-case-015-lsp.yamlconfig2.3 KB
templates/harness-eval-cases/run-local-eval.mjsscript43.3 KB
templates/harness-eval-cases/run-underspec-n3.mjsscript41.1 KB
templates/harness-eval-external/deepswe/README.mdmarkdown9.9 KB
templates/harness-eval-external/deepswe/prepare-supergoal-arm.mjsscript4.4 KB
templates/harness-eval-external/deepswe/run-default-suite.mjsscript6.1 KB
templates/harness-eval-external/deepswe/run-full-cycle.mjsscript25.3 KB
templates/harness-eval-external/deepswe/task-set.yamlconfig12.2 KB
templates/harness-eval-external/featadd/README.mdmarkdown3.9 KB
templates/harness-eval-external/featadd/features.tsvother1.6 KB
templates/harness-eval-external/featadd/gen.shscript4.0 KB
templates/harness-eval-external/featadd/package.pyscript6.2 KB
templates/harness-eval-gate.mjsscript24.9 KB
templates/harness-eval-report.mdmarkdown5.8 KB
templates/harness-eval-result.jsonconfig11.9 KB
templates/harness-eval-runner.mjsscript19.9 KB
templates/harness-eval-stats.mjsscript1.8 KB
templates/learn-grounding-gate.mjsscript6.1 KB
templates/observability/heartbeat.schema.jsonconfig2.5 KB
templates/observability/sg-emit.shexecutablescript5.4 KB
templates/qa-gate.shexecutablescript6.9 KB
templates/qa-only-gate.shexecutablescript5.1 KB
templates/qa-report.mdmarkdown2.8 KB
templates/rules.mdmarkdown301 B
templates/run-state-gate.mjsscript2.2 KB
templates/run-state.jsonconfig474 B
templates/skill-frontmatter-gate.mjsscript3.7 KB
templates/skill-install-audit.mjsexecutablescript3.7 KB
templates/skill-mine/mine.mjsscript7.8 KB
templates/skill.md.templateother568 B
templates/teach-lesson-gate.mjsscript5.4 KB
templates/teach/README.mdmarkdown4.1 KB
templates/teach/assets/lesson-book.jsscript5.8 KB
templates/teach/assets/lesson-template.htmlother6.1 KB
templates/teach/assets/lesson.cssother11.8 KB
templates/teach/assets/quiz.jsscript3.2 KB

reference/qa-only.md

View on GitHub
# QA-ONLY mode - exercise + verify a running app, write no code

Use when the user wants QA, data verification, or a data comparison and explicitly NOT a code change:
"QA만", "qa only", "검증만 해줘", "데이터 정합성 확인", "API 수정 전후 확인", "A/B 테스트", "데이터 비교".

QA-ONLY drives an already-running app plus optional read-only DB, builds an Impact Matrix, reports what
worked / didn't / was discovered, and can persist a reusable QA suite. It writes NO production code, runs
NO build, creates NO worktree, and skips default-loop phases. It is read-only except the run folder and
the `.domain-agent/qa/` suite.

Fix/feature request -> route to DEBUG/LEGACY; QA-ONLY reports findings and stops.

## Pipeline

`Intake -> Target & Access -> Impact Matrix -> Scenario checkpoint -> Exercise -> Cross-check -> Audit -> Report -> Persist`

| Phase | Goal | Writes | Exit gate |
|---|---|---|---|
| Intake | Capture QA goal, comparison type, acceptance, supplied change/release context | `brief.md`, `state.json` | goal + comparison type stated |
| Target & Access | Resolve running-app target (URL/env, NOT built from repo), browser driver, read-only DB access, action budget | `brief.md` | target reachable; driver + DB access named/skipped; `action_cap` set (default 100) |
| Impact Matrix | Infer the feature's direct, adjacent, before/during/after, data, role, viewport, and failure-risk surface | `brief.md` | coverage plan has priority tiers and explicit exclusions |
| Scenario checkpoint | Show the detailed scenario list + budget + comparison type; let the user narrow | run note | user confirms or narrows; proceed unless told to wait |
| Exercise | Dispatch `qa-tester` to drive the app and produce execution evidence, capped | `QA.md` `## QA`, `qa/`, `state.json` | each scenario observation + evidence path recorded; combined `action_count` written to `state.json`, `<= action_cap` |
| Cross-check | Read-only DB checks via `db-reader`: auth, source-of-truth expected values, dataset/env diff | `QA.md` `## QA`, `qa/` | each check pass/fail + small diff recorded; no raw dumps |
| Audit | Audit — dispatch `qa-auditor` to independently reconcile brief, matrix, ledger, tester/DB evidence, and rerun non-browser checks | `QA.md` verdict | gaps and contradictions named; final verdict is evidence-bound |
| Report | Auditor writes the human-friendly result with coverage, reproduction steps, and remaining risk | `report.md` | required report headings present |
| Persist | Save reusable suite, index it | `.domain-agent/qa/<suite>.md`, `index.md` | suite + re-run steps saved; no secrets/PII; path gitignored |

## Target & Access

- **Target.** Test an existing running app. Ask for URL/env. Start a local server only if asked; never
  build product code to get one. Static single HTML opens via `file://`.
- **Browser driver.** Use `agent-browser` by default (`reference/agent-browser.md`). Use
  `playwright-cli` only when agent-browser cannot complete reliable QA; record `Tool:` and `Fallback:`
  per `reference/qa.md`. Auth uses native session/state/CDP paths.
- **Navigation map.** Load `.domain-agent/qa/nav-map.md` first to reach gated/popup-heavy screens;
  build it on first entry and correct any drifted rows in place (`reference/qa.md` "Navigation map").
- **DB access.** Read-only, DB-independent (`reference/db-access.md`): fetch test auth, verify UI values
  against source of truth, or run dataset/env diffs.

## Impact Matrix (detailed QA model)

Before scenarios, build an **Impact Matrix** from the request, release note/diff summary, nav-map,
domain pack, screen/API calls, and observed behavior.

Include:

- **Feature under test.** The user/business outcome, not just the route or button label.
- **before/during/after actions.** Before: auth, role, permissions, settings, seed/existing data, entry
  route. During: happy path, edges, validation failure, retry, cancel, duplicate submit, repeated run.
  After: refresh, back/forward, reopen, logout/login, revisit elsewhere, and expected next-state checks.
- **Adjacent surfaces.** Screens, tabs, API calls, DB tables, cache/storage, notifications, exports/files,
  search/list views, counters, totals, and permissions that share the same state.
- **Displayed data.** Check displayed data accuracy and consistency across state propagation paths:
  detail, list/search, counters/totals, dashboards, notifications, exports, related records, and
  persisted source-of-truth where they exist.
- **Feature-specific scenario families.** Pick the family that matches the web feature instead of forcing
  a generic CRUD path:
  - CRUD/content publishing: create -> list/search -> detail -> edit -> refresh/reopen -> delete/archive.
  - Workflow/status transitions: start -> approve/reject/cancel -> downstream queue/list/badge update.
  - Commerce/booking: select -> price/availability -> cart/reservation -> payment/confirmation -> history.
  - Auth/account/settings: login/session -> permission change -> protected routes -> logout/relogin.
  - Search/filter/sort/pagination: query -> filter/sort/page change -> item open -> back preserves state.
  - Upload/download/export: upload -> preview/metadata -> persisted listing -> download/export content.
  - Realtime/notification: trigger event -> toast/badge/feed -> refresh/reconnect consistency.
- **Complexity probes.** Include complex multi-step scenarios such as create -> edit -> save -> reopen,
  partial failure -> retry, role switch, concurrent/repeated action, stale-cache refresh, empty data,
  high/low boundary values, mobile/desktop viewport, and keyboard/a11y navigation when relevant.
- **Risk tier.** Mark scenarios `must`/`should`/`could`; run `must` first, then highest-risk `should`
  until the action cap. Never mark an unrun item as covered.

If access/data/fixture/role/session/DB/budget blocks a scenario, list it as uncovered with reason and
residual risk. Do not collapse a detailed QA request into a happy-path smoke.

## Comparison types

State one in `brief.md` `Comparison:`; it picks what the Cross-check phase compares.

| `Comparison:` | Compares | How |
|---|---|---|
| `functional` | scenario outcome vs expected behavior | drive the Impact Matrix must-paths, edge paths, and adjacent checks; assert each step |
| `data-integrity` | UI value vs DB source of truth | read the value on screen, query the DB, diff |
| `before-after` | a saved baseline suite vs the app now (regression after an API/code change) | re-run a persisted `.domain-agent/qa/<suite>.md`; diff against its baseline |
| `ab` | variant A vs variant B (A/B test arms) | run the same scenario on each arm; diff outcome/data |
| `env` | environment A vs environment B | run the same scenario/query on each; diff outcome/data |

## Action cap (anti-context-overload)

Each browser interaction (`open/click/type/fill/snapshot/screenshot`) and DB query counts as one QA
action. Default `action_cap` is **100**. `functional`: all interaction budget goes to `qa-tester`; DB
comparisons reserve a small `db-reader` sub-budget. Auditor evidence review and non-browser command reruns
do not consume browser actions. Agents report counts; conductor writes `state.json.action_count`. At cap,
stop, report done/remaining, and ask before a fresh budget. Gate fails if `action_count > action_cap`.

## Vault (reduced run folder)

`docs/changelog/<YYYY-MM>/<DD-qa-topic>/`: `brief.md`, `QA.md` (`## QA` evidence),
`report.md`, `qa/scenario-ledger.md`, `qa/` evidence, `state.json`. No `PLAN.md`. Per-run report stays
here; reusable suite goes to the domain pack.

## Report (the one human-facing deliverable)

`qa-auditor` writes `report.md` from `templates/qa-report.md` in the docs language (SKILL.md) under
English anchor headings. Include verdict, coverage, per-scenario pass/fail + evidence path, findings, reproduction
steps, not-covered risk, and re-run command/suite. For issues, include URL/env, role/account type (not
credentials), starting state, exact clicks/inputs, expected vs actual, evidence path, and retry count if
intermittent.

## Persist (repeatable + indexed)

Save reusable suites under the domain pack (`reference/domain-context.md` saving loop):

- `.domain-agent/qa/<suite>.md`: Impact Matrix, scenario list, steps, `Comparison:` type, named DB
  checks (no raw secrets/data), saved `before-after` baseline values, reproduction notes, coverage,
  uncovered areas, and residual risks, plus driver-neutral re-run steps. For a Playwright fallback, include
  the spec path and agent-browser limitation.
- Index it in `.domain-agent/index.md` under `## QA Suites` so "re-run the <feature> QA" routes
  instantly.
- `.domain-agent/qa/nav-map.md`: one shared navigation map per repo (entry/auth, routes, popups/new
  tabs, selectors, `screen -> API`). Load it before driving; correct touched drifted rows.
- Same domain-pack rules: read-only repo write, gitignored path, no secrets/tokens/PII/raw rows.

Repeatable QA defaults to saved agent-browser steps. Only a documented Playwright fallback saves
`.domain-agent/qa/<suite>/<flow>.spec.ts` (first run may be hand-driven).

`before-after` re-run: confirm saved baseline date and schema/contract still fit; otherwise re-derive
before diffing.

## Gate

`bash templates/qa-only-gate.sh <vault> <browser|cli>` exits 0 only when: `report.md` carries the
required anchor headings; the underlying `qa-gate.sh` browser/CLI evidence passes; `action_count <=
action_cap` in `state.json`; and, if a DB was used, a `DB:` line marks it read-only and no write verb
(INSERT/UPDATE/DELETE/DROP/ALTER/TRUNCATE) appears in the recorded DB commands. Never edit the gate to
pass.

## Dispatch

Core QA-ONLY uses two separate read-only subagents: `qa-tester` and `qa-auditor`. `qa-tester` produces
execution evidence; `qa-auditor` owns the independent final verdict. Add `db-reader` as an optional
third evidence role only when DB truth is required:

- `qa-tester` (`agents/qa-tester.md`): drives the running app; screenshots/dumps stay in its context.
- `qa-auditor` (`agents/qa-auditor.md`): consumes tester/DB evidence, reruns non-browser checks, and
  writes the final verdict/report without driving the app or DB.
- `db-reader` (`agents/db-reader.md`): reads the DB read-only; raw rows/secrets stay in its context.

For broad QA, split the Impact Matrix across multiple `qa-tester` instances into **Scenario shards**
only when independent; one `qa-auditor` owns the final verdict: direct flow,
adjacent surface, role/permission, displayed-data propagation, viewport/a11y, error/retry, env/A-B arms.
Do not shard ordered flows, shared mutable state, or scarce test data.

Use the vault as the common communication surface:

- `qa/scenario-ledger.md`: the shared ledger owned by the conductor. It contains the Impact Matrix,
  shard list, owner, status, action count, evidence path, covered/uncovered result, and residual risk.
- `qa/shards/<shard-id>.md`: one file per QA subagent. The subagent writes only its own shard file and
  evidence under `qa/`; it never edits another shard or the shared ledger directly.

Conductor merges tester shards and optional DB evidence into `qa/scenario-ledger.md` and `QA.md`, then
dispatches one fresh `qa-auditor`. The auditor reconciles the complete ledger/evidence and writes the
canonical `QA.md` verdict plus `report.md`; the conductor does not invent or override the verdict.

Handoff is conductor-mediated through the vault, never agent-to-agent. For `data-integrity`, `db-reader`
writes sanitized `qa/expected.md` (`field -> expected value`, no raw rows/secrets/PII); the conductor
passes those small values to `qa-tester` for UI diffing and later exposes the saved evidence to the
auditor. Auth/credentials are the exception: hand off transiently in the `qa-tester` prompt, never
written to any file. For `ab`/`env`/dataset diffs with no UI, `db-reader` returns the diff directly.
The conductor passes compressed evidence summaries to the auditor and must never drive the browser or
DB or decide the verdict itself.

`qa/expected.md` also becomes the saved **baseline** for a later `before-after` re-run.

## Stop conditions

- Target not reachable: report it; do not build product code to conjure one.
- A finding needs a code fix: report it as a finding and recommend switching to DEBUG/LEGACY; do not
  edit product code.
- DB only reachable with a write-capable account: still issue read-only statements only; never write.
- DB truth is load-bearing but access is missing/skipped/unsafe: record `DB evidence: Not covered` under
  `## Not covered`, with the blocker and residual risk.
- Action cap hit: stop, report done + remaining, ask before spending a fresh budget.
- DB credentials or PII would land in a saved file: redact; never persist secrets or raw rows.
- Complex or adjacent scenario cannot be exercised safely: record it under `## Not covered` with the
  concrete blocker and residual risk.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall