The Agent Skills Directory

[SAFE]: No malicious patterns or behaviors were identified in the skill. The logic is consistent with the stated purpose of providing rigorous mathematical proofs and verification.- [PROMPT_INJECTION]: The skill ingests data from local theorem notes and drafts (Ingestion: SKILL.md Step 1), which theoretically presents an indirect prompt injection surface. However, the risk is negligible as the skill lacks network access, requires a human-centric research context, and includes internal checks for mathematical rigor that mitigate typical instruction-override attempts.- [COMMAND_EXECUTION]: The skill employs Grep and Glob for identifying relevant local files and context. These are used as legitimate search tools within the agent's restricted execution environment to support the theorem normalization and dependency mapping phases.

proof-writer