The Agent Skills Directory

[SAFE]: The skill is well-defined and uses specific tools within the defillama namespace to fetch public blockchain and protocol data. There is no evidence of credential harvesting, unauthorized network calls, or command execution.- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external API endpoints which serves as a potential surface for indirect prompt injection. 1. Ingestion points: Data is fetched via defillama:get_events and defillama:get_protocol_metrics tools as described in SKILL.md. 2. Boundary markers: Not present. 3. Capability inventory: The skill is limited to information retrieval and formatting, lacking capabilities for file system modification or shell access. 4. Sanitization: The skill does not explicitly filter or sanitize the returned data. This surface is considered low risk and safe given the specific use case and restricted environment.

risk-assessment