define-contracts

Pass

Audited by Gen Agent Trust Hub on Nov 15, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No instructions were found that attempt to override agent behavior, bypass safety guidelines, or extract system prompts. The language is purely technical and instructional.
  • [DATA_EXFILTRATION]: No hardcoded credentials, sensitive file path access (like .ssh or .aws), or unauthorized network operations were detected. Data operations are localized to the dataframes and contract files provided by the user.
  • [OBFUSCATION]: The skill uses clear, standard Markdown and Python. No Base64 encoding, zero-width characters, homoglyphs, or other obfuscation techniques were found.
  • [REMOTE_CODE_EXECUTION]: The skill relies on the 'pointblank' Python library for its core functionality. It does not perform any 'curl | bash' operations or download/execute scripts from untrusted remote sources.
  • [DYNAMIC_EXECUTION]: While the skill mentions the use of 'lambda' expressions and Polars expressions for data validation (e.g., in the 'conjointly' step), these are standard practices for data engineering. The skill explicitly notes that such executable elements cannot be serialized to YAML, which mitigates risks associated with loading untrusted configuration files.
  • [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface for external contract formats (JSON Schema, Frictionless, dbt, ODCS). However, the 'pointblank' library uses a strict whitelist of allowed validation methods, significantly reducing the risk of malicious instructions being embedded in schema definitions.
  • Ingestion points: 'import_contract()' function in 'SKILL.md'.
  • Boundary markers: None explicitly defined for external files, but library-level parsing acts as a structural boundary.
  • Capability inventory: File read/write (YAML/JSON), data validation via Polars.
  • Sanitization: Method name whitelisting in the 'Contract' constructor prevents arbitrary code execution through contract steps.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 15, 2026, 07:28 AM