The Agent Skills Directory

[SAFE]: The skill implements a documentation and workflow management framework with no detected malicious behaviors, obfuscation, or data exfiltration.
[COMMAND_EXECUTION]: The skill instructs the agent to execute project-specific validation commands (e.g., npm test) defined in the plan files to verify each implementation phase. This is a functional requirement for a development agent and occurs within the local environment context.
[DATA_EXPOSURE]: The agent reads and writes documentation files (specs and plans) and standard configuration files (AGENTS.md, CLAUDE.md) to track progress and maintain context, which is consistent with its intended purpose.

spec-workflow