request-claude-review

Runs Claude Code reviews of task planning docs and resolves review comments.

Updated Apr 17, 2026
One-click install
npx skills add https://github.com/mistakenot/skills --skill request-claude-review-mistakenot
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: request-claude-review
Source: https://github.com/mistakenot/skills/tree/main/plugins/planning-workflow/skills/request-claude-review
Command: npx skills add https://github.com/mistakenot/skills --skill request-claude-review-mistakenot

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Getting a second pair of eyes on task planning documents before execution requires manual copy-pasting and follow-up. This Skill automates sending task docs to Claude Code for review and ensures every comment it leaves gets resolved through the proper append-only workflow. ## Core Features & Use Cases - Automated Review Invocation: Launches Claude Code in print mode against the task folder, handling stdin redirection and logging so the review runs unattended. - Comment Detection and Failure Diagnosis: Counts review comments in both Markdown and HTML planning docs, and surfaces log output when the review fails or returns nothing. - Mandatory Comment Resolution: Invokes the resolve-comments workflow and verifies author replies exist, keeping the append-only review thread intact. - Use Case: After drafting a plan for task 042, run this Skill to have Claude review the docs, then automatically resolve all P1/P2/P3 comments before execution begins. ## Quick Start Ask the agent to request a Claude review for task 042 and resolve any comments it leaves.

Frequently Asked Questions about request-claude-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get Claude Code to review my task planning docs?▼

Run this Skill with a numeric task ID and the repository root. It locates the matching docs/tasks folder, invokes claude in print mode with the /review-task command, and reports the comments left in the docs.

How are review comments resolved after a Claude review?▼

After a successful review, the Skill invokes the resolve-comments workflow in the current agent context. Comments are append-only, so resolutions are added as author replies rather than editing or deleting existing threads.

Does the review support both Markdown and HTML planning docs?▼

Yes. Markdown docs use HTML comment blocks with UNRESOLVED/RESOLVED/REJECTED status tags, while HTML docs use pd-thread and pd-comment elements. The Skill counts and verifies comments in both formats.

Why does the Claude review hang or stall before starting?▼

In print mode with an inherited open stdin, Claude waits several seconds for stdin data. The Skill redirects stdin from /dev/null to give an immediate EOF so the review starts without the stall.

What happens if the Claude review fails or leaves no comments?▼

The Skill checks the claude exit code and comment count, then prints the tail of the review log for diagnosis. It stops and reports the failure to the user instead of proceeding to comment resolution.