writing-pull-requests
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes shell command templates to automate GitHub workflows. Commands use the
ghCLI to fetch repository metadata and authentication tokens, andcurlis used to upload files to GitHub's attachment infrastructure. It also manages an orphan assets branch using standard git operations. - [DATA_EXFILTRATION]: The skill documents a method for uploading local files to
uploads.github.com(a GitHub subdomain). As this targets a well-known service for the purpose of PR media integration, it is a functional feature rather than a malicious exfiltration attempt. - [INDIRECT_PROMPT_INJECTION]: The skill instructions define an attack surface by processing user-provided file paths for upload. Ingestion points: The agent is guided to use user-specified file paths for media placeholders in shell commands. Capability inventory: The agent has capabilities for network POST requests and file system operations. Sanitization: The instructions include warnings about data loss with certain commands but rely on user input for path safety.
Audit Metadata