Gh Repo Cartographer

GH Repo Cartographer

Quick Start

Use the bundled script for deterministic inventory work. From this skill or repository directory:

python scripts\gh_repo_cartographer.py --include-pages --include-releases --output github-repo-map.md --json-output github-repo-map.json

For a fast inventory from a saved repository address cache, use:

python scripts\gh_repo_cartographer.py --repo-address-file data\github-repo-addresses-harzva-just-agent.txt --no-fetch --output github-repo-map.md --json-output github-repo-map.json

To refresh that cache with matched local folders:

python scripts\gh_repo_cartographer.py --repo-address-file data\github-repo-addresses-harzva-just-agent.txt --no-fetch --scan-root D:\study\code --repo-address-output data\github-repo-addresses-harzva-just-agent.txt

The script:

• Reads managed account aliases from gh-account-router when available.
• Resolves each alias to its canonical GitHub login with gh api user.
• Lists repositories owned by each resolved login, with a REST API fallback when gh repo list GraphQL calls time out.
• Reads repository URLs from --repo-address-file text caches; when used without --account, this skips live repository enumeration and uses the file as the remote source.
• Resolves router URL aliases such as https://github.com/just-agent, saihao to the canonical GitHub login.
• Optionally fetches GitHub Pages status/URL and latest release metadata for every repository.
• Scans local Git repositories under configured roots.
• Matches local remotes to github.com/owner/repo.
• Fetches remotes by default, then compares HEAD with @{u} to report synced, behind, ahead, diverged, no-upstream, or no-local-copy.

Common Options

• Add scan roots with repeated --scan-root <path>. If omitted, the script uses GH_REPO_CARTOGRAPHER_ROOTS when set, otherwise the current directory.
• Add account aliases with repeated --account <alias>. If omitted, aliases are discovered from gh-account-router.
• Add repository address caches with repeated --repo-address-file <txt>. The checked-in local cache for Harzva and Just-Agent is data\github-repo-addresses-harzva-just-agent.txt.
• Set GH_REPO_CARTOGRAPHER_REPO_ADDRESS_FILE to one or more os.pathsep-separated txt files when a workflow should default to saved repository URLs.
• Set GH_ACCOUNT_ROUTER when gh-account-router is installed somewhere other than ~/.codex/skills/gh-account-router/scripts/gh_account_router.py.
• Use --no-fetch only when the user explicitly wants a faster offline check; mark results as based on stale local remote refs.
• Use --include-pages when building a Pages matrix or homepage preview index.
• Use --include-releases when building a release hub or project matrix with release badges.
• Use --max-depth <n> to widen or narrow local scanning.
• Use --output <file> for a Markdown report and --json-output <file> for downstream README automation.
• Use --repo-address-output <file> to write a URL cache that includes matched local folder paths as local=<path> on each repository line.

Reporting Guidance

Summarize the Markdown report in the final answer:

• Number of remote repositories found.
• Number of repositories with GitHub Pages.
• Number of repositories with releases.
• Number matched to local folders.
• Repositories that are behind, ahead, diverged, dirty, or missing local copies.
• Local repositories with GitHub remotes outside the managed accounts.
• The report paths created.

Never print raw tokens or access-file contents. If authentication fails, report the account alias/login and the GitHub error after token redaction.

README Optimization Workflow

After generating the map, use it to prioritize README work:

1. Start with repositories that have local matches and are synced.
2. For behind or diverged repos, ask before editing README content unless the user clearly asked to update local branches.
3. For no-local-copy repos, recommend cloning before README optimization.
4. Use the JSON report when batching README audits or generating dashboards.