The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified in the instructions or templates. The skill establishes a professional workflow for software development planning without introducing dangerous behaviors.\n- [PROMPT_INJECTION]: The instructions are clear and instructional, focusing on task decomposition and quality assurance. No attempts to bypass safety filters, override system-level instructions, or extract system prompts were found.\n- [COMMAND_EXECUTION]: The skill defines templates for shell commands (such as pytest and git commit) to be included in implementation plans. These are provided as purely instructional examples for the agent to follow and do not involve autonomous or unsafe execution of commands by the skill itself.\n- [DATA_EXFILTRATION]: The skill does not perform any network operations, hardcode credentials, or access sensitive file paths (e.g., .ssh, .env). The specified plan storage location (docs/tmp/) is a standard project-specific directory.

writing-plans