Agent Claude Code
Adversarially reviews a pull request or local diff. Assumes the code is broken until proven otherwise and systematically tries to break it across logic, error handling, security, concurrency, data integrity, resource management, and API contract dimensions. Input: a PR number/URL, or nothing (reviews the local…
Agent Claude Code
Adversarially checks whether an implementation actually delivers what a cockpit issue asked for. Maps the original request and every acceptance criterion against the real diff and tests, then flags unmet criteria, scope gaps, and out-of-scope over-reach. Run by implement-issue first, before the code and UI reviews, as…
Agent Claude Code
Adversarially reviews a cockpit issue plan or draft. Assumes the plan is flawed and that an implementation agent following it will get stuck or build the wrong thing, then proves specific failure modes. Input: a cockpit issue key (e.g. CK-12), or raw markdown of a plan draft.
Agent Claude Code
Reviews UI changes in a cockpit issue's implementation by driving the running app with Playwright. Spins up an isolated cockpit dev server from the implementation worktree, navigates to the affected screens at desktop and mobile viewports, exercises the key interactions, screenshots them, assesses the change against…
Skill Claude CodeCodex
Merge an accepted cockpit issue and clean up its review surface. Picks up an issue in Accepted, confirms the PR is mergeable and its CI is green (bringing the branch up to date with next, and resolving conflicts, only when needed), merges it, kills the test server, removes the worktree, and moves the issue to Done.…
Skill Claude CodeCodex
Verify a cockpit UI/frontend change in a real browser with Playwright — screenshots, mobile viewport, reproduce a visual bug. Use when asked to visually confirm a UI change, screenshot the running app, or reproduce a rendering/layout bug. Covers cockpit-specific setup the generic run/verify skills miss: isolated dev…
Skill Claude CodeCodex
Implement a refined cockpit issue. Branches off next in an isolated git worktree, writes the code following the approved plan, runs build/lint/tests, runs the completeness-reviewer agent first to confirm every acceptance criterion has implementing code, self-reviews with the code-reviewer agent (up to 4 rounds), runs…
Skill Claude CodeCodex
Write an end-to-end integration test that exercises a runtime path against the REAL Claude Code CLI driven by a mock Anthropic API. Use to prove a behavioural acceptance criterion actually works (session spawn, agent tools, permission flow, scheduled jobs, model selection, system-prompt injection) rather than that the…
Skill Claude CodeCodex
Refine a cockpit issue into an exact implementation plan an agent can execute. Reads the target issue, investigates the real code, reproduces bugs, then overwrites the issue description with a precise build spec (files, symbols, exact changes, acceptance criteria). Use when asked to refine, plan, or work up a cockpit…
Skill Claude CodeCodex
Regenerate the CLI slash-command classification map (src/lib/cli-commands.ts) from the installed Claude CLI. Use after upgrading Claude Code, or when a slash command behaves wrong in a PTY session (hangs on "processing" or is wrongly blocked).
Skill Claude CodeCodex
Run an on-demand adversarial review of a refined cockpit issue plan. Dispatches the plan-reviewer agent against the plan in the issue description and posts the findings as a comment. Use when a human wants a fresh review of an existing plan outside the autonomous refinement flow, e.g. "review CK-12". Not part of the…