synthesis-kb-edit

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill implements strong defensive measures, such as verifying that all file operations are contained within the repository root using the relative_to method and using yaml.safe_load for configuration files.
  • [COMMAND_EXECUTION]: The skill performs expected Git operations and executes local validation scripts (kb_config.py, okf_validate.py, okf_consistency.py). These actions are the core functionality and are restricted by the repository's .agents/knowledge-base.yaml policy.
  • [PROMPT_INJECTION]: The skill processes Markdown files, which represents an indirect prompt injection surface. This risk is managed through a mandatory validation chain. 1. Ingestion points: files in the configured bundle_path. 2. Boundary markers: absent. 3. Capability inventory: Git operations, local script execution, and file writing. 4. Sanitization: enforced schema validation, consistency checks, and confidentiality scanning.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 05:31 PM
Security Audit — agent-trust-hub — synthesis-kb-edit