playbooks

home / skills / desplega-ai / ai-toolbox / process-review

process-review skill

/cc-plugin/file-review/skills/process-review

This skill processes file-review comments, extracts HTML markers, guides Claude through addressing each item interactively, and removes resolved markers.

npx playbooks add skill desplega-ai/ai-toolbox --skill process-review

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
393 B
---
name: process-review
description: Process review comments in a file after user finishes reviewing in file-review GUI. Extracts HTML comment markers, guides Claude through addressing each comment interactively, and removes resolved markers.
---

# Process Review Comments

Thin redirect. Follow the **Process Comments** section of the `file-review:file-review` skill for full instructions.

Overview

This skill processes review comments embedded in a file after a user finishes reviewing in a file-review GUI. It locates HTML-style comment markers, walks an AI assistant (Claude) through addressing each comment interactively, and removes markers when issues are resolved. The skill is written for TypeScript-based agent toolchains and designed for seamless integration into an agent workflow.

How this skill works

The skill scans the file for HTML comment markers used by the GUI to flag review notes. For each comment it extracts context and the comment body, then prompts Claude to propose fixes, explanations, or follow-up questions. When a comment is resolved, the skill removes or updates the original marker so the file reflects the current review state. It also emits structured metadata so other agents or tools can track progress.

When to use it

  • After completing a manual review session in the file-review GUI
  • When you want an AI to transform review notes into concrete code edits or explanations
  • To convert unresolved comments into interactive tasks for an assistant
  • When you need automated cleanup of resolved comment markers in source files

Best practices

  • Keep GUI comment markers in a consistent, documented HTML format so the scanner reliably detects them
  • Provide sufficient code context around comments to enable accurate AI suggestions
  • Make single-purpose comments to avoid ambiguity during automated addressing
  • Review proposed AI edits before finalizing to ensure intent is preserved
  • Use structured metadata output to integrate with CI or task trackers

Example use cases

  • A reviewer leaves inline HTML comment markers; the agent processes them into suggested code edits and removes markers after acceptance
  • An agent summarizes multiple review comments into a prioritized action list for developers
  • Interactive loop where Claude asks clarifying questions about ambiguous comments and the user confirms before changes are applied
  • Automated cleanup step in a pull request workflow that removes resolved review markers and adds a resolution note

FAQ

What kinds of comment markers are supported?

The skill expects HTML-style comment markers used by the file-review GUI; keep marker format consistent for reliable extraction.

Does the skill apply edits automatically?

It proposes and can apply edits, but workflows should include a confirmation step so humans can validate changes before finalizing.