AgentStack
SKILL verified MIT Self-run

Github Maintainer

skill-orkas-ai-orkas-awesome-agentskills-github-maintainer · by Orkas-AI

A Claude skill from Orkas-AI/Orkas-Awesome-AgentSkills.

No reviews yet
0 installs
16 views
0.0% view→install

Install

$ agentstack add skill-orkas-ai-orkas-awesome-agentskills-github-maintainer

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

Are you the author of Github Maintainer? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

GitHub Maintainer

Use this when the user wants maintainer-grade judgment over a GitHub project queue. The goal is not just to list issues or pull requests; it is to decide what each item means, what evidence exists, what is risky, and what should happen next.

Scope

  • Default to the current repository when the working directory is a GitHub checkout.
  • Broaden to multiple repositories only when the user asks for broad, all, everything, a named owner, or a named organization.
  • Read local project guidance first when available: CONTRIBUTING.md, README.md, maintainer notes, roadmap, project policy files, or equivalent docs.
  • Use live GitHub data through gh when queue state matters.
  • Only comment, close, merge, rerun CI, or push changes after explicit user approval.

Workflow

  1. Identify the repository and confirm access.
  2. Collect open issues and pull requests.
  3. Inspect enough detail to explain each surfaced item.
  4. Classify each item by type, fit, risk, proof, blocker, and next action.
  5. Include trust signals for non-maintainer contributors when recommending action.
  6. If the user asks to act, work one item at a time and verify before moving on.

References

  • Read references/issue-pr-triage.md for queue discovery, item evaluation, trust signals, and output structure.
  • Read references/permissions-and-safety.md before any write action or autonomous project-maintenance work.
  • Use scripts/github-activity.sh when gh and jq are available and contributor trust history is useful.

Default Output

For current-repository triage:

Repo: owner/name
Source: gh commands and local files inspected

Immediate:
- #123 PR: title
  What: one-line plain-language summary.
  Type/Fit/Risk: bug|feature|dependency|docs|internal; good|mixed|poor; low|medium|high because ...
  Trust: factual contributor signal, or unavailable.
  Proof: CI, tests, reproduction, diff review, or missing proof.
  Blocker: none, missing access, failing check, unclear product direction, stale branch, no repro, etc.
  Next: exact maintainer action.

Needs judgment:
- #124 issue: ...

Defer/close:
- #125 issue: ...

Skipped:
- Reason and scope limitation.

For broad scans:

Owners scanned: owner/org list
Source: repo discovery method and selected gh inspections

Top queues:
- owner/repo: issue count, PR count, why it matters, next action

Immediate actions:
- Specific low-risk item with evidence.

Needs judgment:
- Ambiguous or high-risk item with decision needed.

Skipped:
- Archived, forks, missing access, or out-of-scope repositories.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.