third-party-integrations
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: The skill consists of instructional documentation and best practices for securely building third-party integrations. It does not contain any executable code, obfuscation, or malicious patterns. All instructions promote secure coding and least-privilege principles.- [INDIRECT_PROMPT_INJECTION]: The skill identifies and provides mitigation for the processing of untrusted data from third-party providers.
- Ingestion points: Described in AGENT-ACCESS.md and SKILL.md as webhook bodies, email text, CRM record fields, and form fields.
- Boundary markers: AGENT-ACCESS.md explicitly defines external content as 'data' and 'never instructions', creating a conceptual boundary for the agent.
- Capability inventory: AGENT-ACCESS.md describes agent capabilities to call APIs, perform writes, and use tools based on user intent.
- Sanitization: AGENT-ACCESS.md contains explicit instructions to never interpolate untrusted data into shells, queries, or paths, and to validate all inputs at the boundary.- [EXTERNAL_DOWNLOADS]: The skill contains references to external documentation and examples on GitHub.
- Evidence: References include the author's own repository (github.com/Quegenx/agent-skills) and established technical examples (github.com/coreyhaines31/marketingskills) used for educational purposes.
Audit Metadata