The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill is designed to generate and execute Python scripts at runtime to analyze the installed environment. It uses templates in references/probe_templates.md that utilize importlib and inspect to extract symbol signatures and pydoc for documentation. This execution is restricted to the scratch/ directory for traceability, satisfying the skill's primary requirement of providing verifiable API lookups against the installed version.
[INDIRECT_PROMPT_INJECTION]: The skill ingests external data from package docstrings and versioned web documentation (Shape 3). Malicious content embedded in these external sources could potentially influence the agent's behavior. The skill mitigates this by requiring verbatim caching into structured markdown files with specific headers (e.g., ## Signature, ## help()) and documentation source attribution.
[EXTERNAL_DOWNLOADS]: The skill performs web searches and fetches narrative documentation from external URLs. These references target official vendor resources (e.g., skore.probabl.ai) to ensure accuracy for the specific library versions installed in the workspace.

python-api