idea

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes git commit and git push commands to manage the local ideas repository.\n- [COMMAND_EXECUTION]: It uses the GitHub CLI (gh issue create and gh project item-add) to mirror ideas into the user's GitHub project board.\n- [DATA_EXFILTRATION]: Transmits idea summaries to GitHub. This operation targets a well-known service and uses the repository owner's configuration.\n- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes user-supplied text and incorporates it into files and command parameters.\n
  • Ingestion points: User-provided idea descriptions captured via natural language or command prefixes in SKILL.md.\n
  • Boundary markers: Employs markdown blockquotes to encapsulate verbatim user wording in the generated idea.md file.\n
  • Capability inventory: Includes local file system writes, Git repository modification, and GitHub API interactions through the gh tool.\n
  • Sanitization: Relies on the agent's default handling of string interpolation for shell commands and file writing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 04:37 PM
Security Audit — agent-trust-hub — idea