The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project-local markdown files (plans/*.plan.md) and treats their content as a series of instructions for implementation and verification.
Ingestion points: The agent uses Glob and Read tools to access plans/*.plan.md files which can be modified by any collaborator or via pull requests.
Boundary markers: No specific boundary markers or 'ignore' instructions are defined to separate the data in the plan files from the agent's core instructions.
Capability inventory: The skill possesses powerful tools including Bash for command execution, Write/Edit for file modification, and Agent for delegating tasks.
Sanitization: There is no evidence of sanitization or validation of the commands or instructions found within the plan files before they are processed by the agent.
[COMMAND_EXECUTION]: The 'Verify' step of the execution flow explicitly instructs the agent to 'Run the exact commands or checks described' in the task's acceptance criteria. This allows for arbitrary shell command execution based on the content of the markdown plan files.

plan-execution