The Agent Skills Directory

[SAFE]: Comprehensive analysis of the SKILL.md and prompt.md files found no evidence of malicious intent, obfuscation, or unauthorized data access. All referenced SOPs are internal to the platform environment.- [PROMPT_INJECTION]: The skill accepts user-provided strings for analysis in the claim_or_assumption and context fields. Although this is an ingestion point for potentially untrusted data (indirect prompt injection surface), the prompt instructions are strictly analytical and do not incorporate the input into executable commands or sensitive system operations, maintaining a safe risk profile.

socratic-probing