facts-implement
Pass
Audited by Gen Agent Trust Hub on May 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
factsCLI tool to manage and verify specifications. It executes shell commands throughfacts check, which runs validation logic associated with specific requirements. The agent is also instructed to modify these commands usingfacts editif errors are detected in the specification. - [REMOTE_CODE_EXECUTION]: The agent's core instruction is to "Write the code that makes [a requirement] true," which involves generating and modifying executable source code within the project environment.
- [EXTERNAL_DOWNLOADS]: The skill allows the agent to install third-party dependencies required for the implementation of the specified features.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by consuming external data (the fact sheet) as actionable instructions.
- Ingestion points: Data is loaded from the specification via
facts listandfacts checkcommands. - Boundary markers: No specific boundary markers or safety delimiters are used to separate the external specification data from the agent's behavioral instructions.
- Capability inventory: The agent has the ability to write code, modify the file system, execute shell commands, and install packages.
- Sanitization: No sanitization or validation of the specification content is performed before the agent attempts to satisfy the requirements.
Audit Metadata