fix-github-issue
Pass
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill fetches issue descriptions from GitHub using the
gh issue viewcommand. Since issue content is user-controlled external data, an attacker could embed malicious instructions in an issue body to manipulate the agent's behavior during the implementation phase.\n - Ingestion points: Issue titles and bodies fetched via
gh issue view(SKILL.md).\n - Boundary markers: Absent. There are no instructions to the agent to treat the fetched issue content as untrusted data or to ignore embedded commands.\n
- Capability inventory: Implementation of changes involves file system writes and process execution (running tests) in the local environment.\n
- Sanitization: Absent. No validation or filtering of the issue content is performed before processing the implementation steps.\n- [COMMAND_EXECUTION]: The skill workflow involves executing shell commands via the
ghCLI tool to fetch metadata and view issues. The agent then proceeds to modify files and run commands (e.g., tests, linting) based on the content retrieved from the external source.
Audit Metadata