The Agent Skills Directory

[PROMPT_INJECTION]: The skill enables the ingestion of scientific protocols and community comments from the protocols.io API, which introduces a surface for indirect prompt injection.\n
Ingestion points: Content is retrieved through endpoints described in references/protocols_api.md, references/discussions.md, and references/additional_features.md.\n
Boundary markers: No explicit delimiters or instructions to treat external API responses as untrusted content are present in the documentation.\n
Capability inventory: The skill utilizes the requests library for network operations and standard Python open() for local file access (e.g., uploading files to the API).\n
Sanitization: The provided documentation does not include steps for validating or sanitizing retrieved content before processing by the agent.\n- [SAFE]: Authentication procedures use standard OAuth 2.0 and token-based flows. The skill correctly uses placeholders like YOUR_ACCESS_TOKEN and provides clear guidance on secure token management.\n- [SAFE]: All referenced external domains (e.g., protocols.io, apidoc.protocols.io, doi.org) are legitimate services associated with the skill's stated purpose of scientific documentation.\n- [SAFE]: A discrepancy exists between the metadata author ("K-Dense Inc.") and the platform author ("LeonChaoX"), which is noted as a non-malicious metadata inconsistency.

protocolsio-integration