The Agent Skills Directory

[SAFE]: The skill is a text-processing subagent that contains only natural language instructions.
[NO_CODE]: No scripts, binaries, or configuration for executable code are present in the skill files.
[PROMPT_INJECTION]: The skill processes untrusted 'claim_text', which is a surface for indirect prompt injection. Ingestion points: 'claim_text' in prompt.md. Boundary markers: Absent. Capability inventory: No tools or executable scripts are available to this skill. Sanitization: Absent. Result: Risk is negligible as the skill lacks any capabilities to exploit.

claim-parsing