codify
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Executes local development tools such as git and eslint to inspect project history and effective configurations to ensure alignment with existing practices.- [INDIRECT_PROMPT_INJECTION]: Ingests project-level artifacts including source code, documentation, and commit history which are potentially untrusted.
- Ingestion points: SKILL.md directs the agent to read project configuration files (eslint, tsconfig), internal documentation (README, CONTRIBUTING), and code history.
- Boundary markers: No specific delimiters are employed during the scanning of these files to separate data from instructions.
- Capability inventory: The skill possesses file-writing capabilities, command execution, and the ability to define new agent skills.
- Sanitization: All proposed changes are gated by user consent through interactive prompting tools like AskUserQuestion to mitigate the risk of malicious data influencing the agent.
- [DYNAMIC_EXECUTION]: Programmatically generates and writes new SKILL.md files to capture discovered project procedures and build steps directly within the project environment.
Audit Metadata