Plugin Claude Code
Claude Code plugins for rook (github.com/rook/) maintainers: maintainer-grade code review, backlog triage, systemic-PR campaigns, and rook house conventions.
Instructions file CodexOpenCode
Instructions for jhoblitt/rook-claude, covering agents.md, skill workflows are documented in the readme, reference skills are exempt, what an exempt skill owes instead and diagram conventions.
Plugin Claude Code
Maintainer-grade skills for rook (github.com/rook/): code review, backlog triage, systemic-PR campaigns, and the rook house conventions they enforce.
Agent
Implements a single well-scoped engineering subtask with file modifications. Use with isolation:worktree for parallel independent changes.
Agent
Single-perspective adversarial attacker for rook (github.com/rook/) design review. Spawned by the rook-code-review skill's proposal mode — and by the pre-PR gate escalating on a major-decision diff — with one hostile mandate such as migration, version skew, security boundary, API evolution, operations, multisite…
Agent
Reviews one rook (github.com/rook/) PR or branch to maintainer standard, following the rook-code-review skill's canon. Spawned by the rook-code-review skill for pre-PR adversarial gates and per-PR fan-out on bulk requests; also usable directly for a single independent review with clean context.
Agent
Metadata-depth triage of a batch of rook (github.com/rook/) issues or PRs for the rook-triage skill — classification, completeness, duplicate/cross-link candidates, label proposals, routing suggestions. Analysis only; never writes to GitHub.
Hook
Runs before the agent uses a tool for WebFetch tool calls, executing webfetch-guard.sh via bash. From jhoblitt/rook-claude.
Hook
Runs when you submit a prompt, before the agent sees it, executing rebase-notice.sh. From jhoblitt/rook-claude.
Skill Claude CodeCodex
Use when reviewing, auditing, or sanity-checking rook (github.com/rook/) code, tests, docs, or workflows — a working tree, branch, commit range, or PR; when checking a branch before opening a PR; when reviewing, critiquing, or stress-testing a rook design proposal, design doc, enhancement doc, or design/ change before…
Skill Claude CodeCodex
Use when authoring or shepherding changes to rook (github.com/rook/) repos — committing, pushing branches, opening or updating PRs, weighing review feedback, backporting, regenerating CRDs or generated code, writing tests or GitHub Actions workflows, watching or retrying CI, or before ANY gh write (comment, label…
Skill Claude CodeCodex
Use when taking over, adopting, or superseding a rook (github.com/rook/) pull request — an author who abandoned it, stopped responding to change requests, or produced an AI-burst PR whose substance is still worth landing. Covers both approaches: editing the existing PR's title and description in place, and opening a…
Skill Claude CodeCodex
Use when applying one systemic change across a rook (github.com/rook/) repo as a series of small, well-contained, independently reviewable PRs — dead-code elimination, lint/staticcheck cleanups, API migrations, renames, import or dependency hygiene. Triggers include: "sweep the rook repo for X", "break this into small…
Skill Claude CodeCodex
Use when triaging rook (github.com/rook/) issues or PRs — grooming or clearing the backlog, labeling or categorizing issues, finding duplicate issues, cross-linking an issue to a PR (or a PR to the issue it fixes), checking what info a bug report is missing, deciding who should review a PR or requesting reviewers, or…
Plugin Claude Code
Claude Code plugins for skill and plugin authors: a four-axis gate that reviews skill, agent, and canon prose the way a maintainer reviews code — duplication and drift, concurrency, token usage, and security.
Instructions file CodexOpenCode
AGENTS.md instructions for jhoblitt/skill-review-claude, covering repository conventions, where the rules live, sanctioned audience renderings and design records.
Plugin Claude Code
Four-axis review gate for Claude Code skill, agent, and plugin prose — drift, concurrency, token usage, and security — enforcing the one-normative-home, shape-matches-the-work, cheapest-sufficient-mechanism, and data-carries-no-authority rules across a pre-PR instruction diff or a whole repo.
Agent
Runs the skill-review concurrency axis over one target — a procedure's control flow matches the dependency structure of the work it describes. Dispatched one-per-axis by the skill-review skill; also usable directly for a concurrency-only pass with clean context.
Agent
Runs the skill-review drift axis over one target — every rule has exactly one normative statement, and every other mention is a pointer to it. Dispatched one-per-axis by the skill-review skill; also usable directly for a drift-only pass with clean context.
Agent
Runs the skill-review security axis over one target — content crossing a procedure's trust boundary carries no authority over what it does. Dispatched one-per-axis by the skill-review skill; also usable directly for a security-only pass with clean context.
Agent
Runs the skill-review token-usage axis over one target — every operation runs on the cheapest mechanism that can do it correctly. Dispatched one-per-axis by the skill-review skill; also usable directly for a token-usage-only pass with clean context.
Skill Claude CodeCodex
Use when preparing to commit, push, or open a pull request (gh pr create) on a branch that changes Claude Code skills, agents, plugins, or marketplace repos, and when reviewing or gating such changes on request — a pre-PR gate on prompt prose across four axes, drift (rules restated, pointers stale, no normative home)…