pr
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the repository, including diff outputs, pull request templates, and documentation files such as README.md and AGENTS.md. This ingestion surface could allow malicious instructions embedded in the repository to influence the agent's behavior during the pull request creation process. * Ingestion points: SKILL.md instructions for reading diffs, templates, and repository documentation. * Boundary markers: There are no explicit delimiters or warnings used to prevent the agent from following instructions embedded in these external files. * Capability inventory: The skill has access to shell command execution, git operations, and GitHub CLI (gh) for creating pull requests and uploading attachments. * Sanitization: The skill incorporates a human review step for the draft pull request and recommends manual inspection of screenshots before upload.
- [COMMAND_EXECUTION]: The skill directs the agent to execute commands and tests documented within the repository to verify changes. This involves running potentially unverified code or CLI tools based on the repository's own instructions.
- [DATA_EXFILTRATION]: The skill uses the GitHub CLI (gh) to upload logs, screenshots, and video recordings to GitHub. While the destination is a well-known and trusted service, there is a risk that sensitive environment data or secrets could be captured and uploaded if the agent documents sensitive areas of the application or system.
Audit Metadata