composio

Pass

Audited by Gen Agent Trust Hub on Nov 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill enables agents to read and act upon data from external SaaS applications, creating a surface for indirect prompt injection where malicious instructions could be embedded in external content.\n
  • Ingestion points: Untrusted data enters the agent context via trigger listeners such as GITHUB_COMMIT_EVENT and GMAIL_NEW_EMAIL documented in SKILL.md and references/auth-and-triggers.md.\n
  • Boundary markers: No specific delimiters or instructions to ignore embedded content within ingested payloads are defined in the provided implementation patterns.\n
  • Capability inventory: The skill grants the agent capabilities to perform write actions, including GITHUB_CREATE_ISSUE and GMAIL_SEND_EMAIL as seen in references/sdk-reference.md.\n
  • Sanitization: No explicit sanitization or validation logic for external payload data is demonstrated in the SDK references.\n- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the Composio SDK and related agent integration packages.\n
  • Evidence: Instructions in SKILL.md direct the user to install composio, composio-claude-agent-sdk, composio-core, and composio-claude via standard package managers.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 16, 2026, 05:44 PM