repo-health
Repo Health
One-command health audit for GitHub repositories. Produces a scored report with prioritized fix recommendations.
Install
claude skill add --from https://github.com/thatrebeccarae/claude-marketing/skills/repo-health
When to Use
- Setting up a new repo and want to make sure nothing is missing
- Periodic health checks on active repositories
- Before open-sourcing a private repo (what needs to be added?)
- Onboarding onto an unfamiliar repo to assess its state
- Pre-release hygiene sweep
How to Use
/repo-health [repo-path] # Full audit with scored report
/repo-health [repo-path] --fix # Audit + generate missing files
Default repo-path is the current working directory if omitted.
Procedure
Execute each step in order. Do not skip steps.
Step 1: Validate Repo Path
- If
repo-pathis provided, resolve to absolute path. If omitted, use current working directory. - Confirm the path exists and is a directory.
- Confirm it contains a
.gitdirectory (is a git repo).
If validation fails:
Error: {repo-path} is not a git repository.
Provide a path to a git repo or run from within one.
STOP.
Step 2: Detect Public/Private
Determine repo visibility:
- Check for
.public-repomarker file at repo root -> public - Check if "public" appears in the remote URL -> public
- Run
gh repo view --json isPrivate -q '.isPrivate'ifghis available -> use result - If none of the above resolve it -> ask the user
Store as repo_visibility (public or private). Public repos are held to a stricter standard in scoring.
Step 3: File Presence Checks
Check for each standard file. Track result as PRESENT or MISSING.
| File | Required (Public) | Required (Private) |
|---|---|---|
LICENSE |
Yes | No |
README.md |
Yes | Yes |
.gitignore |
Yes | Yes |
SECURITY.md |
Yes | No |
CONTRIBUTING.md |
Yes | No |
CODEOWNERS |
Recommended | No |
.editorconfig |
Recommended | Recommended |
.github/pull_request_template.md or .github/PULL_REQUEST_TEMPLATE.md |
Recommended | No |
.github/ISSUE_TEMPLATE/ directory (or .github/ISSUE_TEMPLATE.md) |
Recommended | No |
CHANGELOG.md |
Recommended | No |
CODE_OF_CONDUCT.md |
Recommended (public) | No |
Step 4: GitHub Configuration Checks
Use gh CLI to query repo metadata. If gh is not available, skip this category and note it in the report.
gh repo view --json description,repositoryTopics,hasWikiEnabled,defaultBranchRef,homepageUrl
Check each item. Track as PASS or FAIL.
| Check | Criteria |
|---|---|
| Description set | description is non-empty |
| Topics set | repositoryTopics has at least 3 topics |
| Homepage URL | homepageUrl is set (if project has a site) |
| Default branch name | defaultBranchRef.name is main (not master) |
| Branch protection | gh api repos/{owner}/{repo}/branches/main/protection returns 200 (public repos only) |
| Social preview | gh api repos/{owner}/{repo} and check for custom Open Graph image (informational only) |
Step 5: Documentation Quality
README checks:
- README.md exists (from Step 3)
- README length: < 50 lines = WARN (too short), > 500 lines = INFO (consider splitting)
- Has a project description in the first 5 lines (H1 or bold text)
- Has install/setup instructions (scan for "install", "getting started", "setup", "usage")
- Has a license reference (scan for "license", "LICENSE")
- If the
/github-readmeskill is available, note that a deeper README audit can be run via/github:readme audit
CHANGELOG checks:
- CHANGELOG.md exists (from Step 3)
- If present: follows Keep a Changelog format (scan for
## [version headers) - Most recent entry is within the last 90 days (WARN if stale)
Step 6: Code Hygiene
Stale branches:
git branch -r --merged main | grep -v main | grep -v HEAD
git for-each-ref --sort=committerdate --format='%(refname:short) %(committerdate:relative)' refs/remotes/
- Flag branches with no commits in > 90 days as stale
- Count total stale branches
Tracked secrets/env files:
git ls-files | grep -E '\.env$|\.env\.|credentials|secrets'
- Any
.env,.env.*,credentials.*, orsecrets.*files tracked in git = FAIL
Large binary files:
git ls-files | while read f; do
size=$(wc -c < "$f" 2>/dev/null)
[ "$size" -gt 5242880 ] && echo "$f ($size bytes)"
done
- Files > 5MB tracked in git without Git LFS = WARN
- Check for
.gitattributeswith LFS patterns if large binaries exist
Dependency freshness (if applicable):
- If
package.jsonexists: check forpackage-lock.jsonpresence - If
requirements.txtexists: check if pinned (versions specified) - Do NOT run install commands or modify anything
Step 7: Score and Report
Scoring Categories
| Category | Weight | What It Covers |
|---|---|---|
| Standard Files | 30% | LICENSE, SECURITY.md, CONTRIBUTING.md, CODEOWNERS, .gitignore, .editorconfig, templates |
| GitHub Config | 15% | Description, topics, branch protection, default branch |
| Documentation | 25% | README quality, CHANGELOG presence and freshness |
| Code Hygiene | 20% | Stale branches, tracked secrets, large binaries |
| Community | 10% | CODE_OF_CONDUCT, issue templates, PR templates, CONTRIBUTING.md |
Per-Category Scoring
Each category scores 0-100 based on checks passed within it. See REFERENCE.md for the full rubric.
Letter Grade
| Score | Grade |
|---|---|
| 90-100 | A |
| 80-89 | B |
| 70-79 | C |
| 60-69 | D |
| 0-59 | F |
Report Format
=== Repo Health Report: {repo-name} ===
Visibility: {public | private}
Overall Score: {score}/100 (Grade: {letter})
--- Standard Files (30%) --- Score: {n}/100
[PASS] .gitignore
[PASS] LICENSE (MIT)
[MISS] SECURITY.md
[MISS] CONTRIBUTING.md
[PASS] CODEOWNERS
...
--- GitHub Config (15%) --- Score: {n}/100
[PASS] Description: "One-command repo health audit..."
[FAIL] Topics: none set
[PASS] Default branch: main
[FAIL] Branch protection: not enabled
...
--- Documentation (25%) --- Score: {n}/100
[PASS] README.md (142 lines)
[WARN] No install/setup instructions found in README
[MISS] CHANGELOG.md
...
--- Code Hygiene (20%) --- Score: {n}/100
[PASS] No tracked .env files
[WARN] 3 stale branches (>90 days)
[PASS] No large binary files tracked
...
--- Community (10%) --- Score: {n}/100
[MISS] CODE_OF_CONDUCT.md
[MISS] Issue templates
[PASS] PR template
...
=== Prioritized Fixes ===
1. [HIGH] Add SECURITY.md — required for public repos. See template in REFERENCE.md.
2. [HIGH] Enable branch protection on main — prevents force pushes and requires reviews.
3. [HIGH] Add repository topics — improves discoverability. Suggest: {topic1}, {topic2}, {topic3}.
4. [MED] Add CONTRIBUTING.md — guides external contributors.
5. [MED] Add CHANGELOG.md — tracks releases for users.
6. [LOW] Clean up 3 stale branches.
7. [LOW] Add CODE_OF_CONDUCT.md — signals welcoming community.
Priority levels:
- HIGH: Security, legal (LICENSE/SECURITY.md), or protection issues
- MED: Documentation and community gaps that affect usability
- LOW: Cleanup and nice-to-haves
Step 8: Fix Mode (--fix)
If --fix flag was passed:
- Present the list of MISS and FAIL items that can be auto-generated
- Ask the user to confirm which files to create (show numbered list, accept comma-separated selection or "all")
- For each confirmed file, generate it using sensible defaults:
SECURITY.md: Standard vulnerability reporting templateCONTRIBUTING.md: Fork/branch/PR workflow with code style notesCODE_OF_CONDUCT.md: Contributor Covenant v2.1.editorconfig: Sensible defaults for the detected tech stackCODEOWNERS: Prompt user for owner patterns- Issue/PR templates: YAML form-based templates (see REFERENCE.md)
- Do NOT overwrite existing files
- Run infra/PII scrub on all generated content before writing
Key Principles
- Read-only by default. The audit never modifies files unless
--fixis explicitly passed. - Public repos are held to a higher standard. Private repos skip community and some config checks.
- Prioritize actionable output. Every finding includes a specific fix recommendation.
- No false urgency. Recommended items are scored lower than required items. The report distinguishes between must-fix and nice-to-have.
- Respect existing work. Fix mode never overwrites. It only creates missing files.
- Infra/PII awareness. All generated files pass through the standard PII scrub before writing.
Anti-Patterns -- Do NOT
- Modify any files during a standard audit (no
--fixflag) - Run install commands (
npm install,pip install, etc.) - Push changes or create commits -- the user decides when to commit
- Fabricate repository metadata -- use what
ghandgitreport - Rate repos you do not have local access to -- this is a local audit tool
- Skip the PII scrub when generating files in fix mode
More from thatrebeccarae/claude-marketing
facebook-ads
Meta Ads (Facebook & Instagram) platform expertise. Audit campaigns, audiences, creative strategy, pixel tracking, and CAPI. Use when the user asks about Facebook Ads, Instagram Ads, Meta Ads, social media advertising, Advantage+ campaigns, or Meta pixel/CAPI setup.
14content-creator
Comprehensive content marketing toolkit with brand voice analysis, SEO optimization scripts, content frameworks, social media strategy, and content calendar planning. Use when writing blog posts, creating social media content, analyzing brand voice, optimizing SEO, planning content calendars, or developing content strategy. For deep SEO writing optimization, see the seo-content-writer skill.
13icp-research
Build detailed ideal customer profiles with pain points, objections, buying triggers, and messaging angles. Includes community research to find where ICPs gather online and extract their exact language. Use when researching audiences, creating buyer personas, or developing targeted messaging.
12content-pipeline
End-to-end content creation workflow that orchestrates research, editorial review, and social distribution agents in sequence. Use when the user wants to create, review, and distribute content through a multi-stage pipeline, or says "/content-pipeline".
12content-workflow
End-to-end content creation pipeline from research through editorial review to social distribution. Orchestrates a 3-stage workflow: research, draft/edit, and distribute. Supports blog posts, LinkedIn, Twitter threads, newsletters, and essays.
12brand-voice-guidelines
Develop brand voice, tone matrices, messaging frameworks, and brand book documentation. Use when the user asks about brand voice, tone of voice, brand guidelines, messaging framework, or brand consistency.
12