The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes git branch --show-current to automatically determine the active feature context from the local environment.
[DATA_EXPOSURE]: Reads project-specific metadata from files such as tasks.md, spec.md, and test specifications. These operations are limited to the local repository and are consistent with the skill's primary purpose of tracking feature development.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests untrusted data from project files.
Ingestion points: Reads {feature-path}/tasks.md, {feature-path}/spec.md, and specs/tests/{feature-id}.md in SKILL.md (Step 2).
Boundary markers: None present; the skill treats file content as trusted input for parsing.
Capability inventory: The skill has the ability to write files (work-plan.md) and suggest commands for implementation, but it does not execute these implementation commands automatically.
Sanitization: No specific sanitization or escaping of file content is performed prior to interpolation into the output template.
Assessment: Since the skill's primary purpose is to process these files for project management and the results are presented to the user for review, the severity is minimal.

work-plan