The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes external data from protocols.io, such as protocol descriptions, experimental steps, and user comments, which could be leveraged to perform indirect prompt injection attacks.
Ingestion points: Protocol metadata and content (e.g., GET /protocols/{protocol_id} in references/protocols_api.md), user-contributed comments and threaded replies (e.g., GET /protocols/{protocol_id}/comments in references/discussions.md), and file descriptions or tags in workspaces (e.g., GET /workspaces/{workspace_id}/files/search in references/file_manager.md).
Boundary markers: The skill does not provide explicit boundary markers or instructions to the agent to treat the fetched scientific protocol data as untrusted or separate from its operational logic.
Capability inventory: The skill possesses high-privilege tools including Bash(curl:*), Bash(python:*), Read, Write, and Edit permissions, as defined in the allowed-tools section of SKILL.md.
Sanitization: No sanitization, escaping, or validation mechanisms for the data retrieved from the protocols.io API are documented or implemented in the instructions.

alterlab-protocolsio