What does this skill do, and when should you use it?
This skill focuses on gathering recent repository changes and producing structured output for changelog draft preparation. It follows rules to cite PR numbers, highlight breaking/security issues, and ignore dependency and bot noise. As part of the larger loop-engineering repository, this skill is paired with the 'changelog-drafter' pattern, which runs on a daily or tag cadence and produces drafts that can be handed off to other agents for finalization.
Scans recent merges, pull requests, and commits to identify content relevant for release notes. Produces structured per-item blocks and a scan summary. Enforces citation of PR numbers, explicitly surfaces breaking changes and security issues, and filters out pure dependency and bot noise. The skill is implemented as a standard SKILL.md file that can be invoked by AI coding agents.
- Weekly changelog generation for a project where the user wants a quick compilation of all merged PRs and notable commits.
- Pre-release preparation where a maintainer wants an automated draft highlighting breaking changes and security fixes.
- A development team that wants to skip manual Git history review and feed structured output directly into a changelog drafting process.
- Ensuring all changelog entries reference PR numbers for traceability.
How do you install this skill?
- This skill provides abstract instructions without concrete steps; please add detailed procedures before actual use.
- No mention of sensitive data handling or data flow; ensure no credentials are leaked during use.
- The repository targets English users and may depend on overseas services like GitHub; Chinese users may face access issues, consider network environment.
- Shell / CLI
- Local filesystem
The skill is located in the loop-engineering repository at starters/changelog-drafter/.claude/skills/changelog-scan/. Clone the repository or copy the SKILL.md file to your Claude skills directory (e.g., ~/.claude/skills/changelog-scan/). Ensure the folder name matches the skill name to follow the standard.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/cobusgreyling/loop-engineering.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/starters/changelog-drafter/.claude/skills/changelog-scan" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Scan recent merges and PRs to produce changelog draft output.
After installation, invoke the skill with a prompt like 'Scan recent merges and PRs to produce changelog draft output.' The skill will access Git history and return structured blocks and a summary. Specific input parameters are not documented in SKILL.md, so users should rely on natural language instructions and ensure Git access is configured.
What are this skill's strengths and limitations?
- Automates collection of recent changes, saving manual effort.
- Structured output is suitable for downstream automation.
- Enforces PR number citation for traceability.
- Filters out dependency and bot noise to reduce clutter.
- Output is text-based; may not integrate directly with CI pipelines.
- No explicit support for custom filtering rules.
- Requires a Git repository history; may not work in non-Git environments.
- As part of a larger monorepo, standalone usability is not explicitly documented.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Changelog Scanner this page | 43 · Not recommended | ★ 11k | 3d ago | MIT |
| Changelog Scan Skill | 45 · Not recommended | ★ 11k | 3d ago | MIT |
| Release Notes Drafting Skill | 32 · Not recommended | ★ 11k | 3d ago | MIT |
| CodeRabbit Autofix | 53 · Use with care | ★ 188 | 4d ago | MIT |
| Wiki Changelog Assistant ✓ Microsoft · Official | 51 · Use with care | ★ 2.7k | 3mo ago | MIT |
How did FollowSkills review this skill?
The skill only reads repository metadata and produces a report, without performing modification operations, so least privilege is maintained. However, the skill lacks explicit data-flow transparency or sensitive-data handling. The repository has an MIT license, and the publisher is unverified but that is not a deduction. Reliability and effectiveness are based on static assessment.
The skill instructions are simple but lack detailed execution steps, edge cases, and error handling. Although the repository has CI and tests, they do not cover this skill's key paths. Static review cannot verify actual runtime, so not reaching 10.
The scenario is clear (scan merges/PRs/commits to generate changelog), but boundaries, trigger conditions, and environment fit (such as Chinese language support and network reachability) are missing. The repository overall may have English docs, but this skill file has no Chinese instructions, and may depend on overseas services like GitHub, so deduction.
The skill has a short definition and a few key rules, but lacks installation notes, parameter descriptions, examples, known limitations, and versioning. MIT license exists, but maintenance responsibility and update path are unclear. Partially meets but not comprehensive.
The skill's goal (generating a changelog draft) is clear, but the output format and completeness are not detailed. Static review cannot verify if the output is directly usable, and may still require significant manual rework.
The repository has CI workflows and tests, but not for this skill. The skill itself has no tests or reproducible examples. Static review cannot independently verify the skill's behavior.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →