Skills
Navigate
skills/maikotrindade/awesome-react-native-skills/github/Gen Agent Trust Hub

github

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect prompt injection surface through ingestion of untrusted GitHub metadata.
  • Ingestion points: The skill reads pull request titles, branch names, and status through gh pr commands in SKILL.md and references/stacked-pr-workflow.md.
  • Boundary markers: No delimiters or "ignore embedded instructions" warnings are used when processing external PR content to prevent the agent from obeying instructions embedded in PR titles.
  • Capability inventory: The agent is instructed to execute git and gh CLI commands in SKILL.md and references/stacked-pr-workflow.md, including git push --force-with-lease and gh pr merge.
  • Sanitization: Pull request metadata (such as titles) is interpolated directly into CLI command arguments (e.g., gh pr merge <N> --squash --title "<PR title> (#N)") without escaping or validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 06:10 PM