The Agent Skills Directory

[SAFE]: The skill defines a rigorous implementation planning process that follows software engineering best practices, such as DRY (Don't Repeat Yourself), YAGNI (You Ain't Gonna Need It), and TDD (Test-Driven Development).
[SAFE]: References to external sub-skills including superpowers:subagent-driven-development and superpowers:executing-plans are modular components consistent with the author's ecosystem and represent intended functionality.
[PROMPT_INJECTION]: The skill processes implementation specifications, which creates a surface for indirect prompt injection. This is a standard risk for planning-oriented skills and is mitigated by the skill's rigid structural requirements and manual review prompts.
Ingestion points: User-supplied specifications or requirements documents.
Boundary markers: None explicitly present to distinguish untrusted specification data from agent instructions.
Capability inventory: The skill produces implementation plans intended for execution by tools with file system and shell command access.
Sanitization: No active filtering or escaping of input specification content is implemented.

writing-plans