The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes legitimate CLI tools for environment setup and data management, including uv pip install for package management and lamin init/lamin migrate for database administration. These commands are standard for the tool's scientific purpose.\n- [EXTERNAL_DOWNLOADS]: The skill instructions involve fetching biological ontology data from established public repositories via Bionty and citation information from ArXiv. These are well-known scientific services necessary for the skill's primary function.\n- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing external biological data (DataFrames, AnnData). Ingestion points: Artifact.load() and DataFrameCurator (SKILL.md); Boundary markers: Implemented via ln.Schema definitions (references/annotation-validation.md); Capability inventory: Includes file writes, network operations to S3/GCS, and CLI execution; Sanitization: Provided by curator.validate(), which raises ValidationError on failure. The instructions emphasize structural validation before data registration.\n- [CREDENTIALS_UNSAFE]: The skill promotes secure credential management by instructing users to use workload identities or environment variables (e.g., LAMIN_API_KEY, LAMIN_DB_URL) for authentication with remote databases and cloud storage, rather than hardcoding sensitive secrets.

lamindb