watzup
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses standard version control commands including
git status,git diff, andgit logto analyze the session's work. It also executes a local Node.js script located at.claude/scripts/open-report.cjsto automate opening the generated report for the developer. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from code diffs and commit messages. It mitigates injection risks by explicitly mandating HTML-escaping for all placeholder values in the report and employing a strict Content Security Policy (
default-src 'none') within the HTML template to prevent the execution of malicious scripts. - Ingestion points: Git diffs, commit logs, and prompt ledger files are read from the project environment to determine session activity.
- Boundary markers: The skill relies on programmatic HTML escaping rather than explicit text delimiters to isolate data from instructions in the generated report.
- Capability inventory: The skill possesses file writing capabilities (specifically to
tmp/reports/) and local script execution capabilities (node). - Sanitization: The instructions explicitly define HTML-escaping rules for characters like
&,<,>,", and'to be applied to all placeholders in the report. - [DATA_EXPOSURE]: The skill identifies and reads session-related data (diffs and logs). To prevent this data from being accidentally committed to the repository, it includes a check to ensure the report directory is ignored by
git check-ignorebefore writing any files.
Audit Metadata