The Agent Skills Directory

[SAFE]: The skill operates entirely on local files within the project's documentation directory (docs/).
[DATA_EXPOSURE]: The skill reads docs/PRD.md and docs/PLAN.md. These are standard project files and the access is limited to the defined scope of creating an implementation plan.
[COMMAND_EXECUTION]: While the skill instructions involve exploring the codebase, this is a routine task for a development assistant and does not involve executing arbitrary or high-risk system commands.
[INDIRECT_PROMPT_INJECTION]: The skill ingests content from docs/PRD.md, which is a potential surface for indirect prompt injection. However, the skill includes a mandatory human-in-the-loop 'Quiz' step (Step 6) where the agent must present its proposed breakdown to the user for validation before writing the final plan. This provides a strong mitigation against accidental execution of embedded instructions.

planifie