freshdesk-triage

Pass

Audited by Gen Agent Trust Hub on Jul 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the Bash environment to execute the composio CLI tool for searching tools, fetching ticket schemas, and retrieving ticket data from Freshdesk.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from external sources.
  • Ingestion points: Freshdesk ticket subjects, descriptions, and conversation threads are fetched in Step 4.
  • Boundary markers: The instructions do not define delimiters or specific warnings to ignore instructions embedded within ticket content.
  • Capability inventory: The agent has the ability to execute shell commands via composio execute in Bash.
  • Sanitization: There is no explicit sanitization or validation of the fetched ticket content before it is processed for triage.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 1, 2026, 08:14 PM
Security Audit — agent-trust-hub — freshdesk-triage