The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.Popen in scripts/run_eval.py to invoke the claude CLI for trigger testing and subprocess.run in eval-viewer/generate_review.py to manage local server ports using lsof. These operations are functional requirements for its evaluation features.
[PROMPT_INJECTION]: The skill processes user-supplied test prompts which are then executed by subagents, representing a surface for indirect prompt injection. Ingestion points: Prompts are ingested from evals/evals.json. Boundary markers: No explicit delimiters are used to wrap test prompts during execution. Capability inventory: Subagents used for testing have access to filesystem and terminal tools as defined by the skill being tested. Sanitization: No input validation is performed on test prompts.
[EXTERNAL_DOWNLOADS]: The evaluation viewer loads the SheetJS library from a well-known CDN (cdn.sheetjs.com) to facilitate client-side rendering of spreadsheet files, which is a standard practice for this type of utility.

skill-creator