gh-normalize
Skill0 starsChanged 12 days agoMITView on GitHub
Fills in and standardizes GitHub repo descriptions and topics for one repo or a whole scope (public, private or all), with a current-to-proposed diff and confirmation per repo. Use when the user wants to fix, fill or clean up repo descriptions or topics after an audit.
In this repository8 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- gh-auditSkill
- gh-onboardSkill
- profile-bioSkill
- profile-readmeSkill
- readme-standardSkill
- release-workflowSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 1 is distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Run only when the user explicitly asks.
- Very specific: line 28
If not logged in, ask the user to run gh auth login (scope repo is needed to edit repos) and wait.
- Very specific: line 71
Never write a proposal marked (needs input) without the user's text.
- Very specific: line 95
Every write needs explicit approval.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
gh-normalize
Brings repo description and topics in line with the standards in references/. One repo or a
batch, confirmed per repo.
Run only when the user explicitly asks.
Usage
gh-normalize [--scope public|private|all] [--repo owner/name] [--confirm each|all] [--only desc|topics] [--dry-run]0. Preflight
gh auth status >/dev/null 2>&1 || echo "not logged in"If not logged in, ask the user to run gh auth login (scope repo is needed to edit repos) and
wait.
1. Config
cfg="${GITHUB_MANAGER_HOME:-$HOME/.agents/github-manager}/config.json"
[ -f "$cfg" ] && cat "$cfg" || echo '{}'description.* and topics.* keys override the defaults. Schema:
../gh-audit/references/config.md.
2. Target (first match wins)
--repo owner/name→ one repo--scope→ all repos in that scope, even when run inside a clone:gh repo list <login> [--visibility public|private] --limit 300 --no-archived --source --json name,description,repositoryTopics,primaryLanguage,visibility(forall, omit--visibility; gh accepts only public, private or internal)- current directory's git remote → one repo (
git remote get-url origin) - none → ask the user (one repo or which scope) and wait for an explicit answer
Skip the profile repo (<login>/<login>), forks and archived repos unless the user names them.
3. Analyse (one repo at a time)
Standards: references/description.md and references/tags.md. Pass their absolute paths and the config.
- unless
--only topics: if your environment supports subagents, delegate toagents/description-analyzer.md; otherwise read../../agents/description-analyzer.mdand follow it inline. - unless
--only desc: same foragents/tag-analyzer.md.
Hand prefetched data from step 2 to the analyzers to avoid extra API calls.
4. Confirmation gate
--confirm each(default): showcurrent → proposedper repo; apply, edit or skip on the user's answer.--confirm all: show the full proposal table, then apply all after one explicit approval.--dry-run: report only, no writes.
Never write a proposal marked (needs input) without the user's text.
5. Apply
Pass values as separate arguments; never build the command by pasting text into a shell string.
gh repo edit "<owner>/<repo>" --description "$new_description"
gh repo edit "<owner>/<repo>" --add-topic "a,b,c"
gh repo edit "<owner>/<repo>" --remove-topic "x,y"6. Summary
NORMALIZE scope=<scope>
changed: <n> skipped: <n> errors: <n>
repo desc topics
my-cli ✎ (181→142) +cli
lib-x - -javascriptRules
- Every write needs explicit approval;
--confirm eachis the default. - Touch only the description and topics of the target repos. No other fields, no other repos.
- Proposals come from the analyzers and the standards, built from real repo data. Nothing invented.
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
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.

