refresh-wiki
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by explicitly instructing the agent that the local schema found in
AGENTS.mdandindex.md"always wins" over its own rules. This establishing of external data as authoritative logic could be exploited by an attacker who can modify those files to override the agent's behavior. - Ingestion points: The skill ingests data from
AGENTS.md,index.md, and various notes from the local vault. - Boundary markers: No delimiters or instructions to ignore embedded commands are present in the skill to isolate ingested content.
- Capability inventory: The agent has the capability to read any file in the vault and write to files within the
wikis/directory. - Sanitization: No sanitization, validation, or filtering of the processed content is implemented before it is acted upon by the agent.
- [EXTERNAL_DOWNLOADS]: The skill references documentation and organizational patterns for knowledge bases from a public GitHub Gist URL.
Audit Metadata