The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill instructions direct the agent to generate and execute a Python script (e.g., /tmp/s2_search.py) to interact with the provided s2.py library. This involves dynamic path searching for the library within the user's home directory (e.g., ~/.claude/skills/). This is an intended function for local tool discovery and script-based API interaction.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from the external Semantic Scholar API (paper titles, abstracts, and TLDRs), creating a surface for potential injection.
Ingestion points: API responses are fetched from api.semanticscholar.org via the s2.py helper module, as described in Phase 2 of the search workflow.
Boundary markers: The output formatting functions (format_results, format_details) do not implement explicit delimiters or instructions to ignore embedded instructions within the academic content.
Capability inventory: The skill possesses the ability to perform network requests to academic endpoints and write result files to the local filesystem (e.g., /tmp/s2_results.md).
Sanitization: No sanitization or filtering of the API-provided data is performed before the content is presented to the agent's context.

semanticscholar-skill