wiki-ingest
Pass
Audited by Gen Agent Trust Hub on May 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it ingests and processes untrusted content from the local filesystem.
- Ingestion points: The skill reads files from the
raw/directory, including transcripts, markdown, and files converted from PDF or video (Step 3). - Boundary markers: No explicit delimiters or instructions are provided to the agent to treat the ingested data as untrusted or to ignore embedded instructions within that data.
- Capability inventory: The skill allows the agent to read from and write to the local directory (updating wiki pages, source summaries, and index files) and execute local search queries via the QMD tool.
- Sanitization: No sanitization or filtering of the processed content is specified before the agent acts on it.
- Mitigation: The risk is mitigated by Step 4, which requires the agent to present extracted points and matched pages to the human user and wait for explicit confirmation before proceeding with any file modifications.
Audit Metadata