Skip to content
Back to skills

Metis Review Task

ASecurity

Dispatch the metis-task-reviewer subagent to judge a diff against the task description and plan. Returns a verdict with per-criterion evidence.

  • 3 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added September 5, 2026
ai-agentsgit

Security analysis

A100/100

Pro scans all 3 files and shows the line behind each finding

Scanned September 5, 2026

npx -y skills add gsaranti/pantheon --skill metis-review-task --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Metis Review Task?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Metis Review Task
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/gsaranti-metis-review-task-pantheon/badge)](https://www.skillsdirectory.com/skills/gsaranti-metis-review-task-pantheon)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: metis-review-task
description: Dispatch the metis-task-reviewer subagent to judge a diff against the task description and plan. Returns a verdict with per-criterion evidence.
---

# $metis-review-task

Dispatch the `metis-task-reviewer` subagent to review one implementation against the description and plan it was meant to satisfy.

## Arguments

- **Trailing prompt** — optional. Augments the dispatch prompt per the command-prompts convention (e.g., narrowing the diff range, flagging a specific concern).

## Context expected

The parent collects what the reviewer needs from conversation context before dispatching:

- The **description** the work was implemented against — from this session's chat history, or rehydrated from `.metis/CURRENT.md` via `$metis-session-start`.
- The **plan**, if one was produced — same sources.
- The implementer's **scope report** — produced by `$metis-implement-task` at the close of implementation; in chat history or in `.metis/CURRENT.md`.

If any of these are missing, ask the user before dispatching. Reviewing an unstated brief is the silent-drift mirror of implementing one.

## Preflight

Run this skill's `references/review-task-preflight.sh`. It exits non-zero if the working directory isn't a git repo. On success it reports two fields:

- `BASELINE` — the branch the implementation diverges from. First existing of `main`, `master`, `origin/main`, `origin/master`; empty if none exist (the subagent then reviews only uncommitted changes).
- `DIFF_PRESENT` — `yes` if there are uncommitted changes or commits ahead of `BASELINE`; `no` otherwise.

If `DIFF_PRESENT=no`, surface an "empty diff" finding without dispatching — no implementation work to judge.

## Subagent dispatch

Dispatch the `metis-task-reviewer` subagent with a single structured prompt containing:

- The description.
- The plan, if one was produced.
- The implementer's scope report.
- The `BASELINE` from the preflight (the branch the subagent should diff against).
- Any trailing prompt the user passed to this skill, verbatim.

The subagent reads the git diff itself; the parent does not need to embed it.

## Write scope

This command writes nothing.

## Invocation prompt

Trailing prompt: forwarded verbatim to the subagent.

## Return

Relay the subagent's return:

- **Verdict** — `approve`, `approve-with-nits`, or `reject-with-reasons`.
- **Per-criterion results** — pass/fail with evidence, for each criterion the description and plan implied.
- **Scope reduction findings** — whether the implementer's scope report honestly describes what's missing, and any reductions the implementer didn't surface.
- **Regression findings** — existing behavior the diff broke or put at risk outside the task's own criteria; a confirmed regression drives `reject-with-reasons`.
- **Code-quality notes** — separate from spec compliance.
- **Next step**:
  - `approve` / `approve-with-nits` — user commits/merges; `$metis-session-end` then moves the task to *Built* and advances the active-task pointer.
  - `reject-with-reasons` — back to `$metis-implement-task` with the reviewer's findings as the trailing prompt.

If the subagent returns an "empty diff" finding instead of a review, pass it through as-is.

Files in this skill

  • SKILL.md3.2 KB
  • agents/openai.yaml43 B
  • references/review-task-preflight.sh1.7 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.