explain-diff-html
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and summarize untrusted data in the form of code changes, diffs, and pull requests.
- Ingestion points: Reads contents of diffs, branches, and PRs provided by the user or external repositories.
- Boundary markers: None. There are no instructions for the agent to use delimiters or to ignore potential instructions embedded within the code changes being analyzed.
- Capability inventory: The skill uses file-read capabilities to "broadly explore surrounding code" and file-write capabilities to save an HTML file to a global path (e.g., /tmp/).
- Sanitization: The instructions do not include requirements to escape or sanitize the code content before interpolating it into the final HTML/JavaScript report, potentially allowing the input data to influence the agent's output logic or execute script in the user's browser.
- [DYNAMIC_EXECUTION]: The skill instructs the agent to generate a self-contained HTML file containing CSS and JavaScript to create interactive quizzes and diagrams. While the intended use is for reporting, generating executable code (JavaScript) based on external, untrusted code diffs without sanitization carries a risk of cross-site scripting (XSS) or script injection if the agent is tricked into including malicious logic from the PR into the generated file.
Audit Metadata