The Agent Skills Directory

[SAFE]: The skill incorporates explicit security principles, instructing the agent to never weaken security controls for performance and to prioritize security constraints over speed gains.- [SAFE]: All shell commands described are standard Go development workflows, such as benchmarking and profiling, and do not involve untrusted remote resources or suspicious exfiltration patterns.- [SAFE]: The skill promotes the use of official Go security tools like govulncheck and gosec to verify code integrity and check for vulnerabilities.- [SAFE]: The skill includes a security checklist specifically designed to prevent common issues such as path traversal, credential leakage, and unsafe concurrency patterns.- [SAFE]: While the skill processes untrusted Go source code (representing an indirect prompt injection surface), the analysis is confined to performance and security metrics using standard tooling rather than interpretive execution of instructions within the code. Ingestion points: Go source files and repository metadata read via Read and Glob tools. Boundary markers: Absent. Capability inventory: Execution of shell commands via Bash for tests and profiling. Sanitization: Handled by standard Go compiler and toolchain.

go-optimizer