make-pr
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEDATA_EXFILTRATIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DATA_EXFILTRATION]: The skill captures screenshots of the application UI, which may include sensitive PII or session data if the user follows instructions to log in before capture. These images are pushed to a remote branch in the repository.
- [DYNAMIC_EXECUTION]: The skill generates shell scripts (e.g.,
.scratch/<feature>/screenshots.sh) intended for future execution. These scripts are constructed using feature names and paths inferred from the repository structure and external documentation (PRDs). - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the repository (commit messages, Architecture Decision Records, and PRDs) to generate the PR body and instructions. Inadequate sanitization of this content could lead to the generation of misleading PR descriptions or malicious script commands.
- [COMMAND_EXECUTION]: The skill performs multiple shell operations using
git, the GitHub CLI (gh), and a specializedagent-browsertool to manage branches, commit history, and browser automation.
Audit Metadata