workflow-template-extractor
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The workflow directs the agent to execute project-defined scripts such as
install,dev, andbuildto verify the resulting template. This involves running potentially complex shell commands dictated by the content of the source project. - [INDIRECT_PROMPT_INJECTION]: The skill operates on untrusted external data in the form of an entire project repository. It lacks safeguards to prevent the agent from being influenced by instructions or malicious scripts embedded within the processed project files.
- Ingestion points:
source_repo_root(files from the project being templated) andextract_spec.md(extraction instructions). - Boundary markers: The skill does not define delimiters or specific instructions for the agent to ignore prompt-like content within the source files.
- Capability inventory: The agent performs file system writes and copies, and executes shell commands for verification.
- Sanitization: Partial sanitization exists for secret removal and de-branding, but there is no filtering of source content to prevent command injection or instruction overrides during the templating process.
Audit Metadata