tsg-flow-contract-creator

Pass

Audited by Gen Agent Trust Hub on Oct 3, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from PRD files to generate technical contracts.
  • Ingestion points: tasks/prd-[slug]/prd.md, project baselines, and ADRs.
  • Boundary markers: No explicit prompt delimiters or safety instructions are defined for processing external PRD content.
  • Capability inventory: The skill can write files and execute shell commands for validation.
  • Sanitization: No input sanitization or validation of the PRD content is mentioned.
  • Multi-step chains: Skill output is passed to tsg-flow-techspec-creator for further implementation mapping.
  • [DYNAMIC_EXECUTION]: The skill provides a Python script template in references/odcs.md to be executed via uv run. This uses a heredoc to pass code directly to the Python interpreter for contract validation.
  • [COMMAND_EXECUTION]: The skill instructions mandate the execution of shell commands for validating generated contracts.
  • Commands include spectral, asyncapi validate, and uv run.
  • These commands use computed paths based on the [slug] variable.
  • [EXTERNAL_DOWNLOADS]: The skill relies on external resources and tools for its core functionality.
  • Fetches the official ODCS JSON schema from the standard's repository on GitHub.
  • Executes remote packages using npx for OpenAPI and AsyncAPI validation.
  • Installs Python dependencies (jsonschema, pyyaml) at runtime via uv run.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 3, 2026, 05:26 PM