The Agent Skills Directory

[SAFE]: No signs of malicious instructions, credential harvesting, or unauthorized network activity were found in the skill's metadata or instructions.
[NO_CODE]: The skill is entirely instruction-based and does not include any code files or scripts, eliminating risks associated with code execution.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes document content provided by the user.
Ingestion points: Untrusted document text is read into the context via the Read tool.
Boundary markers: The instructions do not define clear delimiters or include warnings to help the sub-agent ignore instructions embedded within the documents.
Capability inventory: The skill is configured to use the Agent tool for task delegation and the Read tool for accessing user files.
Sanitization: No evidence of input filtering or content sanitization is present before document data is passed to the sub-agent.

review-doc