ia-writing
Pass
Audited by Gen Agent Trust Hub on Jul 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of Markdown files containing writing principles and prohibited phrase lists. It does not perform any network operations or system modifications. The technical specification (SPEC.md) includes specific guidelines to prevent the storage of credentials or sensitive filesystem paths.
- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it is intended to process untrusted user-provided prose.
- Ingestion points: The skill receives external text for editing, rewriting, or humanizing as described in SKILL.md.
- Boundary markers: It utilizes a structured 'Two-Phase Audit Workflow' (references/audit-workflow.md) to separate detection from correction, which serves to isolate analysis from text generation.
- Capability inventory: The skill possesses no access to sensitive tools, subprocess execution, or network capabilities across its instructions and references.
- Sanitization: The audit process requires the agent to quote specific snippets and tag them, providing a structural verification step against unintended instruction following from input text.
Audit Metadata