The Agent Skills Directory

[SAFE]: The skill acts as a documentation generator. It takes project requirements and transforms them into a structured implementation plan following a specific markdown template.
[COMMAND_EXECUTION]: The skill instructions specify shell commands (e.g., pytest, git commit) to be included as text within the generated plan documents. These commands are intended for documentation and are not executed by the skill itself during the planning process.
[DATA_EXPOSURE]: File system access is limited to saving planning documents to a specific subdirectory (docs/plans/). There is no evidence of the skill accessing sensitive files, credentials, or environment variables.
[REMOTE_CODE_EXECUTION]: The skill references other internal capabilities (e.g., superpowers:executing-plans, superpowers:subagent-driven-development). These appear to be platform-specific skill references and do not involve downloading or executing code from external, untrusted sources.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied specifications to generate output.
Ingestion points: Accepts project specifications or requirements as input (SKILL.md).
Boundary markers: No explicit delimiters are used to separate user input from the planning instructions.
Capability inventory: The skill is limited to writing markdown files to the local directory.
Sanitization: No explicit sanitization or validation of the input specification is performed.
Risk Assessment: Because the skill's output is purely documentation-focused (Markdown files), the risk of malicious instructions being executed at the planning stage is negligible.

writing-plans