The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill connects to the official STRING database API (string-db.org) to fetch protein-protein interaction networks and functional enrichment data. This is a well-known bioinformatics resource and is a legitimate destination for the tool's scientific functionality.
[COMMAND_EXECUTION]: Executes a Python CLI wrapper (scripts/string_cli.py) via uv run to manage API requests and data processing. The script uses structured parameter passing and standard library functions to interact with the external service.
[INDIRECT_PROMPT_INJECTION]: The skill processes biological data returned from the STRING API, which could theoretically contain instructions. It mitigates this risk by providing specific instructions to avoid reading large outputs into the agent's context and recommending external tools for data processing.
Ingestion points: External interaction and annotation data received from API endpoints at string-db.org.
Boundary markers: Rule 2 in SKILL.md explicitly instructs the agent to process large outputs using external utilities like jq or grep rather than reading them directly into the context window.
Capability inventory: The tool has capabilities to perform network POST requests to the STRING API and write binary content to user-specified output files.
Sanitization: API results are handled as binary data or structured text (TSV) and written to files, preventing the raw data from being interpreted as instructions by the agent.

string-database