kb-lint

Pass

Audited by Gen Agent Trust Hub on Aug 13, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local Node.js script located at scripts/conformance.mjs to perform deterministic metadata inventory and conformance checks on the knowledge bundle. This is a standard architectural pattern for such tools.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the target bundle directory. During the 'Drift audit' (judgment phase) and 'Fix mode', malicious instructions embedded in the analyzed markdown or YAML files could attempt to influence the agent's behavior or the modifications it performs.
  • Ingestion points: Markdown files and YAML frontmatter within the user-specified <bundle-dir>.
  • Boundary markers: The instructions do not define explicit delimiters or warnings to separate processed content from instructions during the audit phase.
  • Capability inventory: The skill has read access to the bundle, write/modification access to files (in fix mode), and append-only access to log.md.
  • Sanitization: No specific sanitization, validation, or escaping of the processed data is described before it is evaluated by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 13, 2026, 08:35 AM
Security Audit — agent-trust-hub — kb-lint