github-cli_skill

This skill safely wraps gh commands to perform GitHub operations with risk-based prompts and mandatory confirmations for destructive actions.
  • Python

9

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill georgekhananaev/claude-skills-vault --skill github-cli

  • SKILL.md10.4 KB

Overview

This skill is a safety-first wrapper around the GitHub CLI (gh v2.86+). It classifies every GitHub command by risk level and enforces mandatory confirmations for destructive and forbidden operations. The goal is to let you automate routine GitHub tasks while preventing accidental data loss or dangerous repo changes.

How this skill works

Incoming commands are classified into Safe, Write, Destructive, or Forbidden tiers. Safe commands run immediately, Write commands prompt a brief informational notice, Destructive commands trigger an AskUserQuestion flow for explicit user choice, and Forbidden actions require multi-step typed confirmation and never auto-confirm. The skill also enforces CLI prerequisites and common shell safety patterns before executing gh commands.

When to use it

  • Listing, viewing, or searching PRs, issues, releases, workflows, or repos
  • Creating or editing PRs, issues, releases, or repo metadata (non-visibility)
  • Merging, closing, deleting releases, secrets, or canceling runs (destructive actions)
  • Performing repository transfers, repository deletions, or visibility changes (forbidden)
  • Calling the GitHub API via gh api for GET/POST/DELETE with appropriate confirmation

Best practices

  • Verify gh is installed and authenticated (gh --version, gh auth status) before running commands
  • Respect the risk tier: present the user with clear consequences for Write and Destructive commands
  • Avoid interactive flags and pagers; pipe to cat for large outputs
  • Quote multi-word arguments and use timeouts for long-running operations
  • Never auto-confirm Forbidden operations; require explicit typed confirmations

Example use cases

  • List open PRs and view CI checks for a candidate branch (Safe)
  • Create a release with generated notes after informing the user about assets (Write)
  • Ask the user how to merge a PR (squash, merge commit, rebase) before executing (Destructive)
  • Set or rotate repository secrets, with delete operations gated behind confirmation (Write/Destructive)
  • Prevent accidental repo delete or visibility changes by requiring multi-step confirmation (Forbidden)

FAQ

gh v2.86 or newer is required; verify with gh --version.

What scopes are needed for full functionality?

At minimum grant repo and read:org; gh auth status shows current scopes.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational