The Agent Skills Directory

[COMMAND_EXECUTION]: Utilizes a command-line utility named bd to manage the lifecycle of project tasks (epics, tasks, and dependencies). The usage is strictly confined to organizational tasks and aligns with the skill's primary purpose.
[COMMAND_EXECUTION]: Instructs the agent to provide standard development commands (such as pytest and git) within generated plan documents. These serve as functional templates for the engineer or subagent executing the plan and do not constitute arbitrary execution.
[PROMPT_INJECTION]: The skill ingests untrusted specifications which serves as a potential surface for indirect prompt injection. Ingestion occurs via the input 'spec' or 'requirements'. Boundary markers are absent for this input, but the skill's capability for command execution (via the bd tool) and file system writes are constrained by a highly specific output structure and a mandatory self-review process that acts as a functional sanitization step.

writing-plans