The Agent Skills Directory

[SAFE]: No malicious patterns, such as data exfiltration, credential harvesting, or unauthorized command execution, were found within the skill files.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection because it processes feature specifications provided via a file path. However, the risk is minimized by the skill's narrow scope and clear role definitions.
Ingestion points: Feature specification file path provided as an argument.
Boundary markers: Includes specific constraints such as 'You NEVER write code' and 'Never write code or create files'.
Capability inventory: The skill instructions focus solely on text analysis and planning output; no tools for network access or system modification are requested or used.
Sanitization: Not explicitly detailed in the prompt, but the planning-only constraint serves as a logical barrier.
[COMMAND_EXECUTION]: The skill suggests commands for the user to run (e.g., bundle exec rspec), but it does not instruct the agent to execute these commands autonomously.

feature-plan