The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The scripts/setup_deps.py script installs matplotlib and numpy using pip. These are well-known, standard packages from the official PyPI registry used for the skill's chart-generation features.
[COMMAND_EXECUTION]: The script scripts/setup_deps.py uses subprocess.check_call to manage a Python virtual environment and install dependencies. This is a standard setup procedure for skills requiring specific Python environments.
[COMMAND_EXECUTION]: The test script tests/verify-chart-pipeline.py uses subprocess.run to execute internal Python utility scripts for verifying diagram generation workflows.
[PROMPT_INJECTION]: The skill processes untrusted external data from the web, which constitutes an indirect prompt injection surface.
Ingestion points: Untrusted data enters the agent context via WebSearch, WebFetch, and Playwright browser tools as described in SKILL.md.
Boundary markers: Absent. No specific delimiters or directives to ignore instructions within the fetched web content are defined.
Capability inventory: The skill environment includes subprocess execution (in setup and test scripts) and JavaScript execution in the browser via playwright__browser_evaluate.
Sanitization: Absent. The skill does not explicitly describe filtering or sanitizing content retrieved from external URLs.

deep-research