agentleFS
imbue-aiimbue-ai/sculptor/.claude/skills/address-comments/SKILL.md

address-comments

Skill246 starsChanged 5 months agoMITView on GitHub

Apply CLAUDE:-prefixed comments from a pull request

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 6 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 6

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read the referenced file (if it's an inline comment on a specific line)

    line 20
  2. Distinctive:

    Do NOT push any code

    line 25
  3. Distinctive:

    Do NOT reply to the comments on GitHub

    line 26
  4. Distinctive:

    Only make local changes

    line 27
  5. Very specific:

    If a PR number was provided, use it.

    line 36
  6. Distinctive:

    If it's an inline comment, note the file path and line number for context

    line 46
  7. Distinctive:

    Read the relevant file(s)

    line 47

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Apply PR Comments

Fetch comments from a pull request and apply all comments that are prefixed with "CLAUDE:".

Instructions

  1. Fetch the PR's comments via gh (see "Fetching comments" below). If a PR number is provided, use it. Otherwise, resolve the current branch's PR with gh pr view --json number.
  2. Parse the output to find all comments (including discussion notes and inline code comments) that start with "CLAUDE:" (case-insensitive).
  3. For each CLAUDE: comment found:
    • Read the referenced file (if it's an inline comment on a specific line)
    • Understand what change is being requested
    • Apply the requested change to the local codebase
  4. Important restrictions:
    • Do NOT push any code
    • Do NOT reply to the comments on GitHub
    • Only make local changes

Fetching comments

Fetch the PR's comments — both inline review comments and general discussion — with gh:

  • Inline review comments: gh api --paginate "repos/{owner}/{repo}/pulls/<N>/comments" — each has .body, .path, and .line (or .original_line). gh substitutes {owner}/{repo} for the current repo.
  • General discussion: gh api --paginate "repos/{owner}/{repo}/issues/<N>/comments".

If a PR number was provided, use it; otherwise resolve the current branch's PR with gh pr view --json number.

Important: the comment JSON can be large and may get truncated — save it to a temp file first, then parse with jq in a separate command (not chained with &&) to avoid shell-parsing issues with the jq expression.

From the fetched comments, select those whose body starts with "CLAUDE:" (case-insensitive). For each, capture the body and — for inline comments — the file path and line. GitHub's REST comments have no resolved flag, so rely on the diff to see what's already done.

Applying Changes

For each CLAUDE: comment:

  1. Extract the instruction (everything after "CLAUDE:")
  2. If it's an inline comment, note the file path and line number for context
  3. Read the relevant file(s)
  4. Apply the requested change
  5. Move on to the next comment

Summarize all changes made at the end.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.