consult-info-session
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions include a command to install the 'zynkr-skill-builder' tool from a remote GitHub repository (
https://github.com/peter-tu-zynkr/zynkr-skill-builder). It also references the installation of sibling skills ('consult-transcriber', 'consult-launch-comms') from the same origin. - [COMMAND_EXECUTION]: The skill documentation suggests using
npx skills addto install the necessary components, which executes code from a remote source. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data that could contain malicious instructions, creating a potential injection surface.
- Ingestion points: The skill reads session invitation text from Google Docs (
[Comms]Doc) in Step 1 and ingests transcription data provided by theconsult-transcribertool in Step 5. - Boundary markers: No explicit delimiters or boundary markers are defined for the ingested text in the instructions.
- Capability inventory: The skill possesses capabilities to manage calendar events, create and update Google Drive files, and update CRM deal notes via MCP tools.
- Sanitization: A critical human-in-the-loop mitigation is present in Step 6, where the user must manually approve the recap text before it is filed to Drive or drafted in Gmail.
Audit Metadata