The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses markdown templates to generate shell commands like git commit and pytest. These commands are intended to be executed by the agent or a sub-skill to implement the feature described in the plan.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. 1. Ingestion points: External project specifications and design requirements from users or preceding workflow skills. 2. Boundary markers: Absent; user specifications are processed without explicit delimiters or warnings to ignore embedded instructions. 3. Capability inventory: Local file system writes to docs/plans/ and shell command execution using tools like git and pytest. 4. Sanitization: Absent; input requirements are interpolated directly into generated plan templates without validation or escaping.

task-planning