The Agent Skills Directory

Prompt Injection (LOW): The skill processes untrusted external data (project 'specs' or 'requirements') to generate markdown plans and shell commands. This creates a surface for indirect prompt injection.
Ingestion points: Processes 'spec or requirements' for multi-step tasks.
Boundary markers: Absent; there are no instructions to ignore embedded commands within the input data.
Capability inventory: Generates file-write operations to the docs/plans/ directory and produces shell commands for git and test execution.
Sanitization: No sanitization or validation of the input requirements is defined in the instructions.
Data Exfiltration (SAFE): No hardcoded credentials, sensitive file path access, or unauthorized network operations were detected.
Remote Code Execution (SAFE): The skill does not download external scripts or packages and does not utilize dynamic code execution functions like eval or exec in its own logic.

planning