The Agent Skills Directory

[SAFE]: The skill follows a well-defined and restricted procedure for generating implementation plans (plan.md) within specific thread directories (docs/threads/). It acts as a template-based file generator.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests content from external files like seed.md and decisions.md to draft plans. However, the risk is negligible as the skill does not possess capabilities to execute code or perform network requests.
Ingestion points: The skill reads data from seed.md, decisions.md, and external code/issue references (specified in SKILL.md).
Boundary markers: Absent; the skill does not explicitly use delimiters or instructions to ignore embedded commands when interpolating external content into the generated markdown.
Capability inventory: File system access is limited to reading and writing markdown files within a thread-specific root. No subprocess, dynamic execution, or network tools are utilized.
Sanitization: The skill does not perform sanitization or validation on the external content before drafting the implementation plan.

plan-brief