Knowledge Base Builder

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves ingesting potentially untrusted data, which presents an attack surface for indirect prompt injection.\n
  • Ingestion points: The workflows in SKILL.md explicitly instruct the agent to extract knowledge from 'existing documentation', 'code repositories', 'team conversations', and external web sources such as 'Stack Overflow' and 'blog posts'.\n
  • Boundary markers: There are no instructions or templates provided that use clear delimiters or warnings to ignore malicious instructions embedded within the data sources.\n
  • Capability inventory: The skill is intended to be used with tools like Memory MCP (memory.create_entities, memory.create_relations) which allow the agent to write to persistent knowledge storage, and it also performs extensive file system reads.\n
  • Sanitization: The skill lacks any guidance or logic for sanitizing, validating, or filtering the content ingested from external or project sources before processing it.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 02:42 AM
Security Audit — agent-trust-hub — Knowledge Base Builder