raise-pr

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by ingesting untrusted data from external issues and interpolating it into instructions without sanitization.\n
  • Ingestion points: Fetches Jira issue summaries via the Jira REST API and GitHub issue titles via the gh CLI (Step 1.5, Step 11).\n
  • Boundary markers: Absent. The skill embeds external metadata directly into pull request body templates and git commit messages.\n
  • Capability inventory: Extensive subprocess execution (git, yarn, gh, curl, acli) and file system access.\n
  • Sanitization: No filtering or escaping is applied to the external issue metadata before interpolation.\n- [COMMAND_EXECUTION]: The skill performs extensive shell command execution for its core functionality, including git repository management and build processes with yarn. It also uses python3 -c to execute inline utility scripts for parsing JSON responses from network APIs (Step 10.2, Step 11).\n- [EXTERNAL_DOWNLOADS]: The workflow involves fetching dependencies via yarn and communicating with well-known services (GitHub, Atlassian/Jira) to manage development tasks, following documented authentication best practices for these platforms.\n- [REMOTE_CODE_EXECUTION]: The skill processes data fetched from official GitHub and Jira API endpoints using inline Python scripts for JSON parsing. These utility operations target well-known services.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 08:39 AM
Security Audit — agent-trust-hub — raise-pr