The Agent Skills Directory

[NO_CODE]: The analyzed skill file (SKILL.md) contains only markdown documentation, tool signatures, and usage instructions. It lacks any executable code (e.g., Python or JavaScript) or installation commands.
[PROMPT_INJECTION]: (Indirect) The tools agent_identity and agent_reputation are designed to ingest content from untrusted external sources like blockchain state and remote metadata URIs. 1. Ingestion points: Blockchain data and remote JSON metadata files. 2. Boundary markers: No delimiters or instructions to disregard potential commands in the external data are specified. 3. Capability inventory: Tools for wallet creation and on-chain identity registration. 4. Sanitization: The documentation provides no information regarding the validation or filtering of external inputs.

agent-identity