mdbase

Pass

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
  • [SAFE]: No malicious patterns or behaviors were identified in the skill instructions. The provided specification includes security-conscious rules such as path traversal checks to ensure file operations remain within the project boundaries.\n- [NO_CODE]: The skill is purely instructional and consists of markdown files (instructions, specification, and adapters). It does not include any executable scripts, binaries, or external dependencies.\n- [PROMPT_INJECTION]: As the skill involves the agent reading and processing markdown files and YAML frontmatter, it presents an indirect prompt injection surface.\n
  • Ingestion points: Local markdown files (records) and type definitions in the _types/ folder.\n
  • Boundary markers: YAML frontmatter is delimited by standard --- markers.\n
  • Capability inventory: The skill defines actions for the agent to create, read, update, and delete local files within the project directory.\n
  • Sanitization: The instructions require the agent to validate all operations against the mdbase schema and ensure YAML syntax is well-formed.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 25, 2026, 06:48 AM