github-ops
Skill1 starsChanged 6 months agoMITView on GitHub
Use when working with remote GitHub repositories, Actions, PRs, issues, releases, or code search via gh. Use local tools for a checkout, the `web-search-router` skill for open-web research, and a browser for GitHub UI.
23 other projects use a version of this.
Adopted by
Most recently taken up by
- Rojios/everything-claude-codetheir own version, first seen Nov 8, 2026
- Lukk17/agent-standardstheir own version, first seen Nov 7, 2026
- pankaj-mahaur/Agent-Daemontheir own version, first seen Nov 6, 2026
- renepf/claude-skillstheir own version, first seen Oct 25, 2026
- ARCASSystems/FounderOStheir own version, first seen Oct 24, 2026
- kuailexiaozixin/githubtheir own version, first seen Oct 24, 2026
Other versions23 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 23
- daymade/claude-code-skills★ 1.5ksame name
- Fmarzochi/EGC★ 61same name
- gongyijie85/dsh-ecc★ 8same name
- ARCASSystems/FounderOS★ 7same name
- claude-world/claude-agent★ 5same name
- chadixearth/graphyloop★ 3same name
- ImpactMojo/ImpactMojo★ 3same name
- bjornjee/skills★ 2same name
- salmandev/everything-gemini★ 2same name
- BlueChonk/BlueChonk-skills★ 1same name
13 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository28 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- add-mcp-serverSkill
- agents-md-initSkill
- agents-md-maintainSkill
- codexSkill
- coding-benchmarksSkill
- context7-docsSkill
- craft-mcpSkill
- deep-code-reviewSkill
- herdr-omp-orchestrationSkill
- herdr-orchestrationSkill
- himalaya-emailSkill
- image-annotatorSkill
- linkup-searchSkill
- mise-manageSkill
- omniroute-opsSkill
- omp-opsSkill
- parallel-searchSkill
- playwriterSkill
- prompt-craftSkill
- session-retroSkill
- skill-createSkill
- skills-manageSkill
- surge-cliSkill
- tavily-searchSkill
- web-search-routerSkill
- workflow-auditSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.
0
Common
Most of the 19 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
Use gh for remote GitHub data.
- Distinctive: line 41
Read commands do not mutate remote GitHub state.
- Very specific: line 109
Write multiline bodies with --body-file or a heredoc.
“Common” is measured: kept by at least 6 in 10 of the 19 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
GitHub Operations via gh CLI
Last verified: 2026-09-03 (gh 2.99.0)
Use gh for remote GitHub data. Use local tools for a checked-out repository
and the web-search-router skill for open-web research. Clone only when many
remote files make it cheaper. gh is authenticated and returns structured
JSON — do not fetch github.com HTML or raw.githubusercontent.com. Use a
browser tool for the github.com UI.
Preflight
gh auth statusonce per session if unsure (nevergh auth tokenor--show-token— no tokens in the transcript). If auth fails, tell the user to rungh auth login. Missing scopes (INSUFFICIENT_SCOPES, common:projectfor Projects V2): report the missing scope and handgh auth refresh -h github.com -s <scope>to the user — it needs interactive browser approval; never run it as an autonomous fix. Only whenghis unavailable or unauthenticated and the data is public may you fall back to unauthenticatedapi.github.comrequests — label the result as a degraded read; never web-scrape private repo pages.- Commands that need repository context (
pr,issue,run,workflow,release) take-R owner/repooutside a checkout.gh repo viewandgh repo clonetakeOWNER/REPOpositionally.gh apitakes no-R: its{owner}/{repo}placeholders resolve from the local checkout orGH_REPO=owner/repo; explicit paths (repos/owner/repo/...) andgh searchneed neither. - For scripted use, prefix commands that might page or prompt:
GH_PAGER=cat GH_PROMPT_DISABLED=1 gh ...
Reads vs writes
Read commands do not mutate remote GitHub state; downloads may write locally. Anything that mutates GitHub state — commenting, reviewing, creating or closing issues, rerunning or dispatching workflows, enabling/disabling workflows, creating releases — is an external side effect visible to other people: run it only when the user's request clearly asks for that specific mutation; otherwise confirm first.
Context economy
Request only the fields needed: --json field1,field2 --jq '...'. Avoid
dumping raw responses. Bare --json lists fields. --jq is embedded; on
normal commands it requires --json, on gh api it works directly.
Task → command map
| Task | Command |
| Read a repo's README | gh repo view owner/repo |
| Read one file from a remote repo | gh api 'repos/owner/repo/contents/PATH?ref=BRANCH' -H 'Accept: application/vnd.github.raw+json' |
| List a remote directory | gh api repos/owner/repo/contents/DIR --jq '.[].path' |
| Search code across GitHub | gh search code QUERY --language go -L 20 |
| Search code in one repo | gh search code QUERY -R owner/repo (checked out locally? use rg instead) |
| Search issues / PRs / repos / commits | gh search issues|prs|repos|commits ... |
| CI status of a PR | gh pr checks N --json name,state,bucket,link |
| Why did CI fail | gh run list -b BRANCH -s failure -L 5 --json databaseId,displayTitle,url then gh run view ID --log-failed |
| Read PR + conversation comments | gh pr view N --json title,body,state,reviewDecision,comments |
| Read inline code-review comments | gh api repos/{owner}/{repo}/pulls/N/comments (NOT available via --json) |
| Comment on a PR / issue | gh pr comment N --body-file f.md / gh issue comment N -b "..." |
| Approve / request changes | gh pr review N --approve / --request-changes -b "..." |
| PR diff | gh pr diff N [--name-only] |
| Issues | gh issue list -R owner/repo -s open --json number,title,labels, gh issue view N --comments |
| Releases | gh release list, gh release view TAG --json body,assets, gh release download TAG -p '*.tgz' |
| Anything else | gh api <REST path> or gh api graphql -f query='...' — see references/api.md |
Without the raw Accept header the contents endpoint returns base64 JSON.
Skip raw.githubusercontent.com (404s on private repos) and download_url
(temporary token). Many files: gh repo clone owner/repo -- --depth 1.
Searching GitHub
Raw qualifiers mix into the query (gh search code panic path:pkg language:go).
Negated qualifiers must go after --: gh search issues -- "crash -label:bug".
--search-type, qualifier cheat sheet, query limits, and search vs list:
references/search.md.
Actions / CI failures
gh pr checks 123 --json name,state,bucket,link # bucket: pass|fail|pending|skipping|cancel
gh run list -b my-branch -s failure -L 5 --json databaseId,displayTitle,url
gh run view <run-id> --log-failed # only the failed steps' logs
gh run view <run-id> --json jobs --jq '.jobs[] | {name, databaseId, conclusion}'
gh run view --job <job-databaseId> --log-failed # narrow to one jobJob IDs, log association, reruns, watch, artifacts, workflow dispatch: references/actions.md.
PR comments and reviews
A PR has three distinct comment surfaces; gh pr view --json only covers two:
| Surface | Read | Write |
| Conversation comments | gh pr view N --json comments | gh pr comment N --body-file f.md |
| Reviews (approve / request changes) | gh pr view N --json reviews,reviewDecision | gh pr review N --approve / --request-changes / --comment -b "..." |
| Inline code comments (diff-anchored) | gh api repos/{owner}/{repo}/pulls/N/comments | reply: gh api -X POST repos/{owner}/{repo}/pulls/N/comments/<id>/replies -f body='...' |
Write multiline bodies with --body-file or a heredoc; \n in -b renders
literally. For idempotent bot comments, use
gh pr comment N --edit-last --create-if-none -F f.md. Inline comments,
thread resolution, and creating reviews:
references/pr-reviews.md.
Done
Done means the target repo/PR/run is identified and only requested fields are returned; a CI diagnosis names the run, job, failing step, and URL/log evidence; auth, scope, and degraded-read failures are stated explicitly.
Sources
- gh manual: https://cli.github.com/manual
- gh api: https://cli.github.com/manual/gh_api
- gh search: https://cli.github.com/manual/gh_search
- GitHub REST search: https://docs.github.com/en/rest/search/search#constructing-a-search-query
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.

