go-to-k

Detect and delete leftover AWS resources from cdkd integration tests. Only targets resources matching known cdkd stack name patterns.

Proactively hunt for cdkd bugs by deploying real CDK apps that exercise common-but-untested AWS resources, configs, and CloudFormation notations against real AWS, then fix what breaks. Use for a periodic "find latent bugs" sweep, not for verifying a specific change.

Work through already-filed GitHub issues (typically the bug-hunt's output) end to end — triage safely, pick a few FILE-DISJOINT issues to fix in parallel, claim each on the issue before starting (collision-safe with other agents), verify against real AWS, then carry each through merge → pull → release → rebuild the linked binary → worktree cleanup. Use when asked to "handle/address filed issues", not to hunt for new bugs (that is /hunt-bugs).

Recommend which integration tests to run, based on the integ ledger (staleness / last result) plus the code areas touched by recent commits. Outputs a prioritized list of `/run-integ <name>` commands. Use before a release, after a batch of merges, or when unsure what integ coverage a change needs.

Scaffold a new integration test for cdkd. Creates a minimal CDK app with the specified AWS resources for deploy/destroy E2E testing.

Run integration tests (deploy + destroy) against real AWS. Use when you need to verify cdkd works end-to-end with actual AWS resources.

Recommend the right reviewer count for a PR based on size + bias factors. Outputs a concrete plan (inline spot-check / 1 reviewer / 3-axis parallel) plus ready-to-paste Agent dispatch prompts when reviewers are warranted.

Comprehensive PR readiness check before merge. Run quality checks, tests, CI, documentation, AWS resource cleanup, and code review.

Proactively hunt for cdk-local bugs by synthesizing real CDK apps that exercise common-but-untested AWS resources, configs, and CloudFormation notations, then RUNNING them locally in Docker (invoke / start-api / run-task / start-service / start-alb / start-cloudfront / invoke-agentcore / studio) and catching local-execution failures + divergences from deployed behavior. Use for a periodic "find latent bugs" sweep, not for verifying a specific change.

Work through already-filed GitHub issues (typically the bug-hunt's output) end to end — triage safely, pick a few FILE-DISJOINT issues to fix in parallel, claim each on the issue before starting (collision-safe with other agents), verify, then carry each through merge (via /merge-pr) → pull → worktree cleanup. Use when asked to "handle/address filed issues", not to hunt for new bugs (that is /hunt-bugs).

Squash-merge a PR and fully clean up the feature worktree + local branch, without tripping the side-worktree `'main' is already used by worktree` fatal and without bypassing the merge-time gates.

Detect and delete leftovers from interrupted cdk-local runs — Docker containers / networks, AND orphaned vitest forks worker processes that failed to terminate (issue

Triggers when a PR adds a new subcommand factory, CLI option, public helper, or behavior change to cdk-local that cdkd (host CLI) may need to inherit. Walks the diff and pins each category to "exported via internal.ts / inside add<Cmd>SpecificOptions / cdkd notified" — sets the cdkd-parity marker so gh pr create can fire.

Check if documentation (README.md, .claude/CLAUDE.md) is up to date with recent code changes. Use when code has been modified and docs may be stale.

Comprehensive PR readiness check before merge. Run quality checks, tests, CI, documentation, Docker / integ verification, and code review.

Recommend the right reviewer count for a PR based on size + bias factors. Outputs a concrete plan (inline spot-check / 1 reviewer / 3-axis parallel) plus ready-to-paste Agent dispatch prompts when reviewers are warranted.

Proactively hunt for cdkrd bugs by deploying real CDK stacks that exercise common-but-untested AWS resources, configs, and CloudFormation notations against real AWS, then catch false positives + missed detection and fix what breaks. Use for a periodic "find latent bugs" sweep, not for verifying a specific change.

Discover and delete leftover cdkrd TEST AWS resources (ephemeral stacks + stack-external orphans like IAM roles, log groups, RETAIN resources), then release the bughunt-clean gate. Use as the cleanup phase of /work-issues live-tests and /hunt-bugs, or standalone to sweep test debris. NEVER touches non-cdkrd or production resources.

Work through already-filed GitHub issues (typically the bug-hunt's output) end to end — triage safely, pick a few FILE-DISJOINT issues to fix in parallel, claim each on the issue before starting (collision-safe with other agents), verify, then carry each through merge → pull → release → global install → worktree cleanup. Use when asked to "handle/address filed issues", not to hunt for new bugs (that is /hunt-bugs).

Run local quality checks (typecheck, lint, build, tests). Quick check during development.

Comprehensive pre-release verification. Run quality checks, docs consistency, a live-test of changed behavior, and a short retrospective before tagging a release.

Check if documentation (README.md, DESIGN.md, docs/) is up to date with recent code changes. Use when code has been modified and docs may be stale.

Run the markgate end-to-end CLI verification script (`.claude/scripts/e2e.sh`). It exercises the full CLI surface — original primitives (set / verify / clear / run / init / version, default key, --hash files + --include, --state-dir, env-var, precedence) plus every feature added in the 2026-05-09 batch (completion, config lint, TTL, --explain, bare status, composes / requires). Wraps the script in `markgate run` so unchanged repos skip the run. Use whenever you want to confirm "all features still work" before declaring done, opening a PR, or merging — or when your changes touch the CLI surface and you want a quick smoke before pushing.

Pre-push / pre-done-declaration discipline for markgate. Covers implementation anti-patterns (no churn, no silent deletion), a widened proactive audit (grep every introduced name, check --help output, smoke the built binary, run make lint — go vet is not enough), and the post-merge manual smoke pattern. Use whenever you're about to say "done", mark a task complete, push a branch, create a PR, or ask for merge. Companion to `iterate-design`, which covers the pre-code design phase.

Pre-code design iteration for non-trivial features on markgate. Forces a design sketch with trade-offs before writing code, and guards against silent opinion-flipping when the user pushes back. Use when the user asks for an opinion ("what do you think?", "should we?", "recommended?"), proposes a new flag / env / config field, or raises a feature with multiple valid shapes. For the matching pre-push / done-declaration discipline (implementation anti-patterns, audit checklist, smoke-the-binary), see the companion `audit-before-done` skill.

Add a new Operator (force-delete support for an AWS resource type that causes DELETE_FAILED, e.g. AWS::Athena::WorkGroup, AWS::IAM::User). Use when adding a new entry to the "Resource Types that can be forced to delete" table. Not for Preprocessors (deletion-protection checks, Lambda VPC detach, etc.) — use `add-preprocessor` for those.

Add a new Preprocessor that runs BEFORE CloudFormation deletion (e.g. deletion-protection check, Lambda VPC detachment). Use when adding an entry to the "Pre-deletion Processing" table. Not for force-deleting resources that cause DELETE_FAILED — use `add-operator` for those.

Orchestrate ecr-scan-verifier integ tests against real AWS. Handles Inspector enable/disable + propagation waits, scan-on-push toggling, image signing (Notation / Cosign KMS / Cosign Public Key / ECR Managed Signing), and cleanup. Use whenever the user wants to run anything under `test/integ/`.

Verify the integ-test skill, test/integ/README.md, and scripts/integ.sh stay consistent. Runs the mechanical script checks PLUS LLM-only semantic checks (mode-description agreement, code-example coherence, cross-reference resolution, depth parity, helper-arg correctness) that a regex can't catch. Sets the `integ-docs` markgate marker on full pass.

AWS CDK のテストを書く / レビューする / 戦略を考える全ての場面で必ず使用する。「CDK のテスト書いて」「このスタックのテスト書きたい」「CDK のテストはどう書けばいい?」「Stack / Construct のテストどうしよう?」などの依頼にも該当する。スナップショット / Fine-grained assertions / バリデーションの 3 種類のうちどれを書くべきか・書かなくて良いかの判断基準とコード例を提供。加えて、機能フラグ(`cdk.json` の context)の実環境との統一や、`NodejsFunction` の esbuild バンドルをスキップしてテストを高速化するなど、**テスト環境セットアップの Tips** も扱う。具体的には `*.test.ts` 編集時、`Template.fromStack` や `aws-cdk-lib/assertions` 使用時、Stack/Construct の単体テスト新規追加時、既存テストレビュー時、テストが遅い / スナップショットが実デプロイと食い違うといった相談時に該当する。