narrative-quality-auditor
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from potentially untrusted sources such as web scrapes, emails, and reviews as documented in the Data Sources section of SKILL.md. However, it incorporates a mandatory evidence chain for safety: 1. Ingestion points include external text and third-party exports; 2. Boundary markers are established by instructions in SKILL.md and references/skill-contract.md to treat input as evidence only; 3. Capability inventory includes potential interactions with marketing platforms listed in references/connectors.md, though no code is provided; 4. Sanitization is handled through explicit directives to ignore embedded instructions in evidence.
- [SAFE]: The skill documentation follows security best practices by requiring explicit user authorization for any persistent writes or external side effects and advising that API keys should be kept out of skill source control. No scripts or external dependencies are included in the skill.
Audit Metadata