メニュー
Scan project once per session. Run first — all other skills depend on this.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Meta-orchestrator — dispatches to the right skill for the task. Use before any non-trivial work.
Library-agnostic Flutter/Dart code review checklist covering widget best practices, state management patterns (BLoC, Riverpod, Provider, GetX, MobX, Signals), Dart idioms, performance, accessibility, security, and clean architecture.
Structured 6-phase debugging. Build feedback loop first, reproduce deterministically, hypothesize with ranked falsifiable theories, instrument one variable at a time, fix with regression test, cleanup. Use when a bug exists, tests fail unexpectedly, or behavior is wrong and cause is unknown.
Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.
Database migration best practices for schema changes, data migrations, rollbacks, and zero-downtime deployments across PostgreSQL, MySQL, and common ORMs (Prisma, Drizzle, Kysely, Django, TypeORM, golang-migrate).
Run integration and e2e tests after unit tests pass. Use after /supergraph:fix when unit tests are green.
| name | scan |
| description | Scan project once per session. Run first — all other skills depend on this. |
| mcp | codebase-memory-mcp |
Load verified, project-scoped graph context once. Follow the shared
references/codebase-memory-contract.md.
Announce: "📡 /supergraph:scan — loading project context..."
Use git branch --show-current, resolve the repository to an absolute path,
and run bin/detect-project.sh when present. Otherwise detect commands from:
| File | Type | Test | Lint | Format | Build |
|---|---|---|---|---|---|
package.json | node | npm test -- | npx eslint . | npx prettier --write . | npm run build |
Cargo.toml | rust | cargo test | cargo clippy | cargo fmt | cargo build |
pyproject.toml | python | pytest | ruff check . | ruff format . | python -m build |
go.mod | go | go test ./... | golangci-lint run | gofmt -w . | go build ./... |
pubspec.yaml | flutter | flutter test | flutter analyze | dart format . | flutter build |
Ask when none match; never guess commands. Derive one stable CBM_PROJECT from
the repository root name and retain it across branches. Default
CBM_INDEX_MODE=moderate.
list_projects and find CBM_PROJECT.index_status(project=CBM_PROJECT).status: "degraded" as requiring a new index.index_repository(repo_path=<absolute repo path>, name=CBM_PROJECT, mode=CBM_INDEX_MODE) when required. Require status: "indexed".index_status; never claim reuse without a healthy response.get_graph_schema(project=CBM_PROJECT) before structural queries.get_architecture(project=CBM_PROJECT, aspects=["overview","layers","boundaries","clusters","hotspots"]).On any mandatory provider/index/schema error: STOP, show the exact error and the
recovery command (codebase-memory-mcp cli index_repository --repo-path <absolute> --name <project> --mode moderate). Do not write freshness state.
Always call Serena initial instructions and activate project when available, then
load a top-level symbols overview. Set SERENA_ACTIVE=true only after success;
otherwise set it false and report Serena unavailable.
.supergraph-envOnly after healthy index_status, schema, and architecture responses:
PROJECT_TYPE=...
TEST_CMD=...
LINT_CMD=...
FORMAT_CMD=...
BUILD_CMD=...
BRANCH=...
GRAPH_PROVIDER=codebase-memory-mcp
CBM_PROJECT=...
CBM_INDEX_MODE=moderate
CBM_INDEXED_AT=YYYY-MM-DDTHH:MM:SS
SERENA_ACTIVE=true|false
SCAN_TIMESTAMP=YYYY-MM-DDTHH:MM:SS
Branch-matched reuse still requires index_status and Serena revalidation.
Report project/type/commands, provider/project/index status, architecture counts available from the response, Serena status, and whether scan was fresh or reused.
project=CBM_PROJECT where supported.