squid

Monitors CI/CD after `git push`. If the pipeline fails, identifies the related task from commit messages, reopens it, fixes the code locally, runs tests, pushes the fix with `Refs

Reads the git diff after a feature is pushed. Tags every finding as Blocker or Nit. Produces ONE rollup task containing all findings. Does NOT read CI, does NOT comment on the PR, does NOT merge. Use after the SWE has pushed the feature branch and before the orchestrator hands the PR back to the human for squash-merge.

Grooms raw tasks into agent-ready specs (acceptance criteria + BDD scenarios) AND does final user-perspective acceptance review after the Tester passes. Use whenever a task needs to be turned into something the SWE can build, or whenever a task needs the final "is this actually right for users?" review before commit.

Implements a single groomed task assigned by the orchestrator. Writes code and tests locally. Does NOT commit until the Tester has reviewed and approved. Use when a task is groomed and ready for implementation, or when the Tester has returned feedback that needs to be addressed.

Reviews the software-engineer's uncommitted work against the spec and acceptance criteria. Runs the full test suite, verifies every acceptance criterion with evidence, and gives PASS or FAIL. Use after the SWE reports an implementation is done, or after the SWE applies fixes for previously-flagged issues.

Periodic architectural sweep of a codebase — reads existing ADRs to avoid re-litigating settled decisions, maps the current module/dependency graph and layering, surfaces 5–10 architectural smells with severity, and emits each finding as a refactor proposal that `/refactor` can consume directly. Trigger when the user says "/architecture-review", "audit the architecture", "what's wrong with this codebase", asks before a major version bump, or when long-horizon tech debt feels untracked.

Create or update a GitHub pull request for the current branch. Use this skill whenever the user says "create PR", "open PR", "update PR", "push and PR", "/create-pr", or when the PR workflow step is reached in the development workflow. Handles both first-time PR creation and updating existing PRs with new changes.

Run a single task through the inner SWE↔Tester loop with the human supervising in real time. SWE implements, Tester verifies (including the e2e adversarial pass), you review the diff and commit. No PM grooming, no PM acceptance, no PR Reviewer, no On-Call. Trigger when the user wants to ship one task under active supervision, or says "/day".

Install Claude Code PreToolUse hooks that block destructive git/shell commands before they execute — force-push to protected branches, `git reset --hard` of unstaged work, `git push --no-verify`, and `rm -rf` paths that escape the worktree. Smart-merges into the project's `.claude/settings.json` and writes one guardian script. Trigger when the user says "/git-guardrails", asks to harden Claude Code against destructive git commands, or is about to run `/night` unattended for the first time.

Adversarially interview a feature spec before plan approval — surface every unresolved decision (auth model, error semantics, persistence, idempotency, observability, rollout, scope edges), draft codebase-grounded answers where possible, and ask the human only the questions that genuinely need their input. Output is a decision-resolved spec ready for PM grooming. Trigger when the user says "/grill-me", asks "is this spec ready", "poke holes in this", "interview me on this plan", or before invoking `/night` on a vague feature description.