skills/paulrberg/dot-agents/yeet/Gen Agent Trust Hub

yeet

Pass

Audited by Gen Agent Trust Hub on Apr 10, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes GitHub CLI (gh) and git commands to manage pull requests, issues, and discussions. It correctly implements single-quoted HEREDOC syntax ('EOF') to safely handle multi-line user-provided content, preventing shell variable expansion and potential command injection. It also runs a local shell script, scripts/get-macos-version.sh, to retrieve system information for environment reporting.\n- [DATA_EXFILTRATION]: For specialized bug reporting workflows targeting projects like Claude Code and Codex CLI, the skill collects environment metadata including the macOS marketing name, terminal identifiers ($TERM_PROGRAM), shell type ($SHELL), and tool versions. This data is gathered specifically for the purpose of populating debugging sections in GitHub issues submitted to official repositories.\n- [PROMPT_INJECTION]: The skill performs semantic analysis on output from git diff and git log to summarize changes for pull request and issue descriptions. This involves processing untrusted repository data; however, the skill is designed to produce summaries for user-reviewed submissions, and it uses secure shell patterns to mitigate technical injection risks.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 10, 2026, 09:34 AM