The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches a brand stylesheet from Vercel's official domain to provide visual styling for the generated artifacts. This is a safe dependency from a well-known service.
[COMMAND_EXECUTION]: The skill utilizes the bun runtime to execute a local script for generating HTML files and uses the open command to present the resulting artifact to the user.
[INDIRECT_PROMPT_INJECTION]: The skill processes external evidence such as test outputs and traces to generate comparison reports, which constitutes an ingestion surface. However, the risk is minimized through sanitization.
Ingestion points: Processes external data including traces, fixtures, and test output to populate report content.
Boundary markers: Basic HTML escaping is implemented for metadata fields (title, subject, summary) in the generation script.
Capability inventory: The skill can execute local scripts via bun, write files using writeFileSync, perform network requests via fetch, and open files using open.
Sanitization: Employs HTML entity encoding for user-provided metadata strings to prevent injection into the generated report structure.

before-after