teams-meeting-pipeline

Warn

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: MEDIUMPERSISTENCEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PERSISTENCE]: The skill describes methods to establish automated tasks for Microsoft Graph subscription renewal using hermes cron add, systemd timers, or crontabs. These mechanisms maintain execution across sessions to circumvent the 72-hour expiration of Graph webhooks.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process Microsoft Teams meeting transcripts, which are untrusted external data sources.\n
  • Ingestion points: Meeting data is retrieved using the hermes teams-pipeline fetch and hermes teams-pipeline run subcommands described in SKILL.md.\n
  • Boundary markers: There are no explicit delimiters or instructions for the agent to ignore potentially malicious commands embedded within the transcripts.\n
  • Capability inventory: The hermes CLI allows for network interactions via the Microsoft Graph API and local file system access for inspecting and running pipeline jobs.\n
  • Sanitization: The skill does not mention any sanitization or validation of meeting transcripts before they are processed for summaries or action items.\n- [COMMAND_EXECUTION]: The skill uses various hermes CLI subcommands to manage authentication tokens, network subscriptions, and meeting jobs, which requires the agent to execute shell commands using identifiers and URLs provided during the meeting ingestion process.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 28, 2026, 12:48 PM
Security Audit — agent-trust-hub — teams-meeting-pipeline