The Agent Skills Directory

[DATA_EXFILTRATION]: The skill is designed to transmit task-specific data (claims, evidence excerpts, and goals) to an external service via the omh_jev_ask tool. This is a primary function that includes mandatory user notification and consent requirements before any data leaves the machine.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted evidence data which could potentially contain malicious instructions. The risk is minimized by the skill's reliance on structured tool calls with fixed presets and an automated process to remove credential-like strings before analysis. Evidence Chain: (1) Ingestion point: evidence_excerpt in SKILL.md. (2) Boundary markers: Present via structured tool preset. (3) Capability inventory: omh_jev_ask tool call. (4) Sanitization: Removal of credential_like_content as described in references/full-contract.md.
[SAFE]: Analysis of the skill instructions and associated reference files revealed no evidence of malicious patterns, such as obfuscation, unauthorized persistence, or privilege escalation.

omh-jev-done-check