The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were identified in the skill. The instructions are focused on development orchestration and safety protocols.
[PROMPT_INJECTION]: The skill safely handles user input via the $ARGUMENTS variable by encapsulating it in <issues> tags and enforcing a strict evidence-based reasoning protocol.
Ingestion points: User input is accepted through the $ARGUMENTS tag in the SKILL.md file.
Boundary markers: Input is delimited using <issues> and </issues> tags.
Capability inventory: The skill coordinates sub-agents capable of reading/writing files, running tests, and managing git repositories.
Sanitization: The skill mandates 'Debug Mindset' and 'Evidence-Based Reasoning' protocols, requiring file:line proof for all claims, which prevents the agent from executing instructions hidden in the user input.
[COMMAND_EXECUTION]: The skill uses specialized sub-agents for standard development operations (e.g., git, testing). There are no triggers for arbitrary shell commands or unsafe script execution.

fix-parallel