The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest requirements and specifications from external phases (like intent-discovery) and transform them into granular implementation plans. This workflow creates an attack surface where malicious instructions embedded in a design document or specification could be translated into harmful shell commands or code snippets in the final plan.
Ingestion points: Requirements, specifications, and designs produced during the 'intent-discovery' phase or provided by a human partner (SKILL.md).
Boundary markers: The skill does not define explicit delimiters or instructions for the agent to ignore or sanitize commands embedded within the input data.
Capability inventory: The generated plans include arbitrary code snippets and shell commands for testing (pytest) and version control (git) (SKILL.md).
Sanitization: No sanitization or validation logic is specified for the input requirements before they are incorporated into the plan.
[COMMAND_EXECUTION]: The skill explicitly instructs the agent to generate and document shell commands for execution. While the examples provided are for standard development tools such as git and pytest, the framework encourages the creation of executable content based on external input.
[SAFE]: The skill references other components in its ecosystem (e.g., godmode:task-runner, godmode:intent-discovery) as part of a modular workflow. These references are standard for coordinating tasks between specialized agent skills.

task-planning