avoid-ai-writing
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it processes untrusted user-supplied text and has the capability to modify local files.
- Ingestion points: The skill processes external text input in
rewriteanddetectmodes and reads content from local files ineditmode (SKILL.md). - Boundary markers: Explicit instructional boundaries are defined, directing the agent to treat input content strictly as text under audit and to flag rather than follow any instructions found within the input, such as "ignore the rules above" (SKILL.md).
- Capability inventory: The skill has the capability to read and modify local prose files when operating in
editmode (SKILL.md). - Sanitization: The skill relies on instructional boundaries to prevent the agent from obeying injected content; no automated sanitization, escaping, or strict schema validation for the input data is described.
Audit Metadata