glean-common-errors

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides a diagnostic curl command template to verify API connectivity. The command correctly uses an environment variable ($GLEAN_API_KEY) for authentication and targets the official service domain (glean.com), avoiding hardcoded credentials.- [INDIRECT_PROMPT_INJECTION]: The skill provides an error classification handler that processes data from external API responses (specified in SKILL.md under classifyGleanError). This represents a standard surface for indirect prompt injection where untrusted data (error messages) is ingested into the agent's context. Boundary markers and sanitization are absent in the provided logic, but the skill's capabilities are limited to Read, Grep, and Bash for connectivity testing.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 06:23 PM
Security Audit — agent-trust-hub — glean-common-errors