writing-editor
Pass
Audited by Gen Agent Trust Hub on Feb 26, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill processes user-provided drafts for analysis and editing (SKILL.md). It lacks explicit boundary markers or instructions to ignore commands embedded within input text, creating an indirect prompt injection surface. The impact is negligible as the skill lacks access to networks, file systems, or sensitive data.\n
- Ingestion points: Processes written drafts provided for review (SKILL.md).\n
- Boundary markers: Not implemented; input is not wrapped in delimiters.\n
- Capability inventory: Limited to generating structured analysis and line edits; no executable code or privileged access is present.\n
- Sanitization: Not implemented; input is analyzed as provided.\n- [EXTERNAL_DOWNLOADS]: Includes an informational reference link to Wikipedia's guide on AI writing (references/AI_TELLS.md), which is a well-known and trusted service.\n- [NO_CODE]: The skill consists entirely of Markdown documents and contains no scripts, binaries, or executable code.
Audit Metadata