ExtractWisdom
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes the
fabricCLI tool (fabric -y "URL") within theWorkflows/Extract.mdprocess to retrieve transcripts from YouTube URLs.\n- [EXTERNAL_DOWNLOADS]: The skill fetches external data from YouTube and general web articles usingWebFetchandfabricto gather source material for analysis.\n- [PROMPT_INJECTION]: The skill processes untrusted external content, which presents a surface for indirect prompt injection.\n - Ingestion points: YouTube transcripts and article text are ingested into the context during Step 1 of the
Workflows/Extract.mdworkflow.\n - Boundary markers: The skill does not define explicit delimiters or instructions to disregard potential commands hidden within the ingested transcripts or articles.\n
- Capability inventory: The skill has the capability to execute the
fabriccommand and read local files from paths like~/.opencode/and the skill's own directory.\n - Sanitization: No sanitization or filtering logic is specified for the external content before it is processed by the AI.
Audit Metadata