The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data from the protocols.io API, which could potentially contain malicious instructions embedded within protocol steps, descriptions, or comments.
Ingestion points: Data is retrieved through various API endpoints described in references/protocols_api.md, references/discussions.md, and references/file_manager.md, including protocol text, step content, and user-generated comments.
Capability inventory: The skill instructions include network operations (via the Python requests library and curl) and file system interactions (via file downloads and uploads documented in references/file_manager.md and SKILL.md).
Boundary markers: The provided documentation and instructions do not define explicit delimiters or include warnings for the agent to ignore instructions that might be embedded within the processed protocol data.
Sanitization: There are no specific instructions or examples for sanitizing, validating, or escaping external content before it is processed by the agent's language model context.

protocolsio-integration