teams-meeting-pipeline
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
hermes teams-pipelineCLI tool for status checks, meeting job management, and subscription maintenance. These commands are executed locally through the agent's terminal capabilities. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external meeting transcripts from Microsoft Graph. This establishes a vulnerability surface where adversarial content within a meeting transcript could attempt to override agent instructions during the summarization process.
- Ingestion points: External meeting transcripts and notes retrieved via the Microsoft Graph API as referenced in the
fetchandruncommands. - Boundary markers: The skill documentation does not provide specific prompt delimiters or warnings to ignore instructions within the ingested data.
- Capability inventory: Execution of terminal commands and access to the local environment and network APIs.
- Sanitization: No content sanitization or validation steps for ingested meeting artifacts are described in the operational guide.
Audit Metadata