implement-plan
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is designed to execute automated verification steps including build commands, test suites, and linting tools. These are standard operations for a development orchestrator but involve the execution of shell commands that may be influenced by the contents of the implementation plan.
- [INDIRECT_PROMPT_INJECTION]: The skill processes implementation plans stored in the
.humanlayer/tasks/directory. These files contain the instructions that guide the agent's actions and the commands it executes. If the source of these plan files is not fully trusted, they could serve as a vector for malicious instructions. - Ingestion points: Implementation plan files (e.g.,
NN-plan-DESCRIPTION.md) located within the.humanlayer/tasks/directory. - Boundary markers: The instructions do not specify any delimiters or safety prompts to prevent the agent from following malicious instructions embedded within the plan files.
- Capability inventory: The skill has the ability to invoke subagents, execute shell commands (for builds and tests), perform file system operations via
ls, and commit changes to a Git repository. - Sanitization: There is no evidence of sanitization or validation of the plan file content before it is used to direct the agent's behavior.
Audit Metadata