GitHub Reader
Uses the gh CLI (which calls the GitHub REST/GraphQL API) for everything. The CLI handles auth, pagination, and rate limiting; AUP exempts API access from its scraping clauses.
Prerequisite
gh auth status
If not logged in: gh auth login.
Pick the right reference
Identify what kind of GitHub URL the user supplied, then follow the matching reference file:
| URL shape | Reference |
|---|---|
github.com/{owner}/{repo} (root, README, file tree) |
references/repo.md |
github.com/{owner}/{repo}/issues/... |
references/issues.md |
github.com/{owner}/{repo}/pull/... |
references/pulls.md |
github.com/{owner}/{repo}/discussions/... |
references/discussions.md |
github.com/{owner}/{repo}/blob/..., tree/..., code search |
references/code.md |
github.com/{owner}/{repo}/releases/... |
references/releases.md |
For Gist URLs (gist.github.com/...), use the readable-gist skill instead.
General tips
- Prefer
ghsubcommands (gh repo,gh issue,gh pr, ...) where they exist; fall back togh apifor endpoints they don't cover. - Use
--json {fields}to keep responses small. Each subcommand documents its--jsonfield set; rungh {sub} view --helpto list them. - Rate limit: 5,000 req/hour for authenticated users.
gh api rate_limitreports current usage.