The Agent Skills Directory

[SAFE]: No malicious behavior or intentional security bypasses were detected in the skill instructions or metadata. The skill focuses on standard performance engineering practices.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute profiling utilities and benchmark tests. This is a standard requirement for performance engineering and is managed through the skill's optimization workflow.
[PROMPT_INJECTION]: The skill demonstrates an indirect prompt injection surface as it reads and modifies external code files using powerful tools.
Ingestion points: Source code files read via Read, Grep, and Glob tools.
Boundary markers: Absent; the instructions do not specify delimiters for distinguishing code data from system instructions.
Capability inventory: Access to Bash for command execution and Write/Edit for file system modification.
Sanitization: None described for input code content.
Mitigation: The skill's non-negotiable requirement for adversarial-verify (regression testing) provides a security control that prevents the acceptance of optimizations that change intended behavior, thereby mitigating functional impact from potential injections.

auto-performance