Agent
Implements a whole small PR test-first in one agent — plans behavior slices, drives RED→GREEN→refactor for each through the public interface, follows the project rules, and lands one commit. Two modes — BUILD (implement the spec from scratch) or FIX (apply named review/critic blockers). Used by the make-pr-lite skill.
Agent
Scores the comments and docstrings a diff adds against comments.md — one score, a verdict, and one finding per comment that must be cut, shortened, moved, rewritten, or added, each carrying the replacement text. Judges only comments; the reviewer owns code quality and the critic owns goal-fit. Used by the make-pr…
Agent
Judges goal-fit — whether a change actually accomplishes its stated task, not whether the code is well-written. Scores 1-100, gives a verdict, lists what's missing. Distinct from the reviewer (code quality). Used by the make-pr architect and the make-pr-lite skills before declaring a task/PR done.
Agent
Reviews a diff for code quality, correctness/bugs, and security against the project rules. Reports structured findings with severity; does not fix code. Answers "is this code good?" — distinct from the critic, which answers "did it achieve the task?". Used by the make-pr architect and the make-pr-lite skills.
Agent
Writes exactly ONE failing test for ONE behavior through the public interface, runs it, and confirms it fails for the right reason (the RED step). Never writes production code. Used by the architect skill at the start of each TDD cycle.
Agent
Implements exactly one scoped task in one of three modes the architect names — GREEN (minimal code to pass one named failing test), REFACTOR (behavior-preserving restructure, tests stay green), or NON-BEHAVIORAL (an exact config/rename/docs edit) — then commits. Scope-locked: never exceeds the single task it was…
Agent
Frozen fixture agent for installer e2e snapshot scenarios.
Skill Claude CodeCodex
Second frozen fixture skill for installer e2e multi-skill scenarios.
Skill Claude CodeCodex
Frozen fixture skill for installer e2e snapshot scenarios.
Skill Claude CodeCodex
Use when the user wants to turn a discussion into a published plan — a short PRD, vertical slices, and a PR breakdown, all in one growing GitHub issue — or invokes /breakdown.
Skill Claude CodeCodex
Use when PRs have landed mid-build and the plan needs re-checking against a PRD issue's goal and re-planning, or invokes /course-correct.
Skill Claude CodeCodex
Use when the user wants to start the unblocked PRs from a design doc, issue, or artifact in parallel — one tmux window, worktree, and coding agent per PR — or invokes /dispatch.
Skill Claude CodeCodex
Use when the user wants a dead-simple, plain-words explanation of a slice or a PR. It answers why we need it, what it is, and what changes. Also use when the user invokes /explain.
Skill Claude CodeCodex
Use when the user wants to audit a codebase or subsystem for module-decomposition improvements — whether to split, merge, extract, move, delete, or deepen a module to cut complexity — or invokes /improve-architecture.
Skill Claude CodeCodex
Use when the user invokes /latest-rebase or says "rebase on latest main". Rebases current branch onto latest remote main.
Skill Claude CodeCodex
Use when explicitly asked to run /make-pr-lite on an already-scoped, low-risk single-module PR. A cheaper alternative to /make-pr — one self-TDD coder, the language gates, then a parallel panel of 3 reviewers plus a comment-reviewer and a critic, squashed to one commit. Not for feature decomposition, multi-module…
Skill Claude CodeCodex
Use when explicitly asked to run the architect workflow on an already-scoped PR (the /make-pr command) — drives one single-module coding task to done via a deterministic TDD loop: plans behavior slices, dispatches the tdd-runner/worker-coder/reviewer/comment-reviewer/critic agents, runs the language gates, and logs…
Skill Claude CodeCodex
Use when the user wants to babysit a PR through the review/CI cycle, or invokes /pr-babysit. Polls for new review comments and CI failures, fixes them, resolves merge conflicts, and loops until the PR is ready to merge.
Skill Claude CodeCodex
Use when a PRD issue's slice plan needs breaking into small pull requests (35 counted lines each, tests excluded) recorded as a table in that same issue and drawn as a private claude.ai PR-map artifact, or invokes /pr-breakdown.
Skill Claude CodeCodex
Use when the user invokes /pr-explain or wants a reader-facing explainer page for a pull request. It is a plain-words page with five chapters. They are: what & why, a walkthrough of the whole diff, every test enumerated, a scored proof, and confirm commands run with their real output. It is published as a private…
Skill Claude CodeCodex
Use when the user wants to create a PR and babysit it until ready to merge, or invokes /pr-make-till-merge.
Skill Claude CodeCodex
Use when the user invokes /pr-make to create or update a pull request. Alias for /make-pr.
Skill Claude CodeCodex
Use when the user wants this repo's prompt files — skills, agents, rules, the CLAUDE.md template — reviewed against five checks with the failures fixed in place, or invokes /prompt-review.
Skill Claude CodeCodex
Use when PRs have merged and the plan records need catching up — ticks every merged-but-unrecorded PR row (✅ + PR link) in the central plan issue, its build-plan artifact, and any cross-referenced issue, or invokes /shipped.