llm-wiki-interview
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions and metadata do not contain any malicious patterns such as obfuscation, remote code execution, or data exfiltration. The operations are limited to reading local wiki files and proposing/creating draft markdown content.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from user transcripts and voice notes, which creates a potential surface for indirect prompt injection.
- Ingestion points: The skill reads from
wiki/index.md,wiki/log.md, and user-provided transcripts or voice-note text. - Boundary markers: No explicit boundary markers or delimiters are defined for the input data.
- Capability inventory: The skill requires read access to the wiki and optional write access for creating draft pages and reports.
- Sanitization: There is no mention of sanitization or filtering for the processed external text.
Audit Metadata