push-skills
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands including
git(status, diff, fetch, merge, worktree, add, commit, push, rev-parse) andgh(run list, run watch, api) to manage repository state and interact with GitHub infrastructure. - [INDIRECT_PROMPT_INJECTION]: The skill processes the contents of local files (e.g.,
AGENTS.md,SKILL.md) and fetches data from the GitHub API. This creates a surface where external content could influence the agent's behavior if it contains instructions. However, the skill explicitly mentions a sanitization pipeline for the public mirror, and its primary purpose is the management of these instruction files. - [DATA_EXPOSURE]: The skill accesses specific files in the user's home directory (
~/.agentsand~/code/AGENTS.md). These paths are application-specific and are managed using standard Git workflows. The skill explicitly notes that private instructions (~/.agents/AGENTS.md) should remain excluded from the public mirror.
Audit Metadata