codekb-ask

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes codebase content and knowledge entries, which creates a surface for indirect prompt injection if malicious instructions are embedded in the retrieved code chunks.
  • Ingestion points: Project code chunks and knowledge entries retrieved via the codekb_ask MCP tool (SKILL.md).
  • Boundary markers: The skill instructions mandate source traceability (verifying sources) and implement a "no-information guard" to prevent the agent from fabricating answers when no relevant data is found.
  • Capability inventory: The skill is restricted to natural language Q&A and context assembly; it does not involve subprocess execution, file system writes, or network exfiltration tools.
  • Sanitization: The instructions do not explicitly detail sanitization or escaping of the retrieved code content, relying on the agent's interpretation of the provided sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 03:54 AM
Security Audit — agent-trust-hub — codekb-ask