agent-stuff

Run a bounded, fail-closed Claude Code review gate over the current git worktree with a configurable Claude model and strict read-only isolation. Use for fresh-context different-family reviews before commit, including requests for Opus, Fable, Sonnet, or an explicitly selected Claude model; drive fix and re-review rounds until the configured reviewer returns CLEAN.

Use before committing an implementation slice when a configurable different-family reviewer should inspect the slice, or when coordinating up to three fix/re-review cycles with Claude, Codex, or Pi. Claude reviews delegate to the supervised claude-review-loop harness.

Use when you want a fresh-context code-review gate before committing — runs Pi only with the approved OpenAI Codex GPT-5.6 Sol model at high reasoning over the current git diff in a bounded, observable loop, and only proceeds when Pi returns CLEAN. Never falls back to Claude/Anthropic, local models, OpenRouter, or another provider. Drives review → fix → re-review up to a round cap. Triggers: "have pi review this", "pi review before commit", "run the pi review loop".

Use before committing an implementation slice when a configurable different-family reviewer should inspect the slice, or when coordinating up to three fix/re-review cycles with Claude, Codex, or Pi. Claude reviews delegate to the supervised claude-review-loop harness.

Use when the user asks for a compact goal prompt, goal condition, create_goal objective, continuation objective, or second-agent goal statement, especially when the destination has a strict character limit such as 4000 characters. Generate a concise, context-specific goal condition from the current worktree, session context, and relevant planning docs without producing a full implementation or review handoff.

Legacy compatibility alias for claude-review-loop. Use only when an existing workflow explicitly invokes opus-review-loop; delegate the review to the canonical configurable Claude gate with the Opus model.

Consume captured pipy parity-loop lessons and materialize each into a gated edit to the skills/docs/tests/harness. Applies only when the working directory is the pipy repo (~/projects/pipy).

Drive one pi-mono parity gap end to end in the pipy repo (select gap, plan, different-family review, implement, docs, review until CLEAN, commit). Applies only when the working directory is the pipy repo (~/projects/pipy).

Verify transient terminal UI behavior with sampled tmux frames. Use when Codex needs to compare CLI/TUI visual states that appear only while output is being generated, such as spinners, loaders, streaming assistant text, cursor/input redraws, footer placement, or pi-vs-pipy terminal chrome parity; especially when final screenshots or final logs may hide timing-dependent layout differences.

Verify transient terminal UI behavior with sampled tmux frames. Use when Codex needs to compare CLI/TUI visual states that appear only while output is being generated, such as spinners, loaders, streaming assistant text, cursor/input redraws, footer placement, or pi-vs-pipy terminal chrome parity; especially when final screenshots or final logs may hide timing-dependent layout differences.

Use when the user asks to wrap up, finalize, commit, sync, promote, merge back, or push a multi-repo workspace, especially ops-control/ops-library workspaces created by just workspace-create. Validate the touched repos, commit each repo separately, run configured workspace sync/promote commands, and report pushed refs without discarding user changes.

This skill should be used when the user asks to "create a goal handoff", "generate a goal prompt", "draft a goal condition", "write a continuation objective", "produce a second-agent goal statement", says "/goal-handoff", or wants a compact goal/objective they can paste into a bounded goal field (such as a 4000-character objective) for another agent session.

This skill should be used when the user asks to "create a review handoff", "generate a review prompt", "hand off for review", "prepare review context", says "/handoff-review", or wants to generate a prompt they can paste into another agent session (Claude Code, Codex, or similar) to perform a code review.

Use when the user asks for a prompt for a second agent to review the current changes, such as "handoff to reviewer", "write me a code review prompt", or similar. Generate a self-contained, context-specific reviewer prompt based on the current worktree, the user request, and relevant specs or docs already known in context, not a generic template.

Generate a context-specific prompt for a second agent to review the current uncommitted changes. Use when the user asks to "handoff to reviewer", "write a review prompt", "prepare a code review for another agent", or similar. Inspects the real worktree - never produces a generic template.

Use when creating or editing standalone Markdown files with Mermaid charts for Marked 2 sharing in both light and dark mode. Apply the proven init pattern, explicit text color, and avoid fragile Mermaid constructs.

Use when the user asks for a prompt for a second agent to implement a change, feature, refactoring slice, or bugfix. Generate a self-contained, context-specific implementation prompt based on the current worktree, requested target, session context, and relevant specs or documentation already known in context.

This skill should be used when the user asks to "create an implementation handoff", "generate an implementation prompt", "hand off for implementation", "prepare implementation context", says "/handoff-impl", or wants to generate a prompt they can paste into another Claude Code session to perform implementation work.

Use when the user asks to commit changes, prepare work for commit, check whether a change is ready to commit, or finalize the current worktree. Inspect the current changes, bring implementation, tests, and relevant specs or documentation into sync if needed, validate the result, and only then commit when appropriate.

Use when creating or editing standalone Markdown files with Mermaid charts for Marked 2 sharing in both light and dark mode. Apply the proven init pattern and avoid fragile Mermaid constructs.

This skill should be used when the user asks to "summarize a YouTube video", "summarize this video", pastes a YouTube URL and asks for a summary, says "what is this video about", or asks to "extract a transcript" from a YouTube link.

Use when the user asks whether current changes are ready to commit, asks for a pre-commit review, or wants commit readiness assessed without actually creating a commit. Inspect the diff, run appropriate checks, and verify tests, docs, specs, and generated artifacts are in sync.

Use when the user asks to commit changes, prepare current work for commit, or assess whether the worktree is ready. Inspect the diff, identify the intended scope, run appropriate checks, ensure docs, specs, and generated artifacts stay in sync, and commit only when the change is coherent and validated.