d6e-plugin-development

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill documents an architecture where AI agents process data from external, potentially untrusted sources, creating a surface for indirect prompt injection.
  • Ingestion points: Workflows utilizing Fetch steps for external APIs and File steps for workspace storage (SKILL.md, references/timeouts.md).
  • Boundary markers: The platform employs ## PLUGIN: header markers to separate plugin instructions, though the documentation does not provide specific guidance on sanitizing dynamic data interpolated into these contexts (SKILL.md).
  • Capability inventory: Plugins can execute arbitrary logic via JS and Docker STFs, perform database operations with sql(), and trigger external side effects through Effect steps (SKILL.md, references/policy-and-instant-run.md).
  • Sanitization: There is no mention of mandatory input validation or sanitization for external content before it is processed by the agent's tools.
  • [EXTERNAL_DOWNLOADS]: The skill describes platform-native tools used to download external files and SaaS content into workspace storage.
  • Evidence: Documentation for d6e_download_external_file and saas-proxy-download tools used for binary data ingestion from SaaS providers (references/saas-and-downloads.md, references/cross-package-recipes.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 12:26 PM
Security Audit — agent-trust-hub — d6e-plugin-development