wiki-ingest-source
Pass
Audited by Gen Agent Trust Hub on Jun 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection surface detected in the data ingestion workflow.
- Ingestion points: The skill ingests raw material from untrusted sources such as web content and paper notes during the source page creation process.
- Boundary markers: The content is written to the markdown body of source files without delimiters or instructions to ignore embedded commands.
- Capability inventory: The skill utilizes file system write capabilities and vault indexing tools which process the ingested data.
- Sanitization: The protocol requires verbatim preservation of content and explicitly prohibits interpretation or synthesis, preventing the sanitization of potentially malicious instructions hidden in the raw material.
Audit Metadata