docs-flag-jargon-skill
Pass
Audited by Gen Agent Trust Hub on May 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious behaviors, obfuscation, or data exfiltration patterns were detected. The skill is restricted to its documented purpose of jargon review using safe local tools (Read, Grep, Glob).
- [SAFE]: The skill ingests untrusted markdown data, creating a surface for potential indirect prompt injection. Ingestion: Step 1 in SKILL.md reads files from $ARGUMENTS. Boundary markers: Not explicitly defined in instructions. Capability inventory: Read, Grep, Glob tools only. Sanitization: None. This surface is evaluated as safe because the skill lacks capabilities for network access, system modification, or code execution.
Audit Metadata