optimize-go-code

Pass

Audited by Gen Agent Trust Hub on Sep 26, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and execute user-provided Go code for profiling and optimization.
  • Ingestion points: The agent reads project source files and benchmarks, as detailed in the SKILL.md workflow.
  • Boundary markers: The workflow requires implementing test oracles to ensure code equivalence, providing a verification step during the optimization process.
  • Capability inventory: The skill executes go build, go test, and go tool pprof commands through the verify.sh utility to perform analysis.
  • Sanitization: Code is executed within the project environment using the local toolchain; no automated code sanitization is performed.
  • [REMOTE_CODE_EXECUTION]: The skill references the execution of a remote Go utility for benchmark analysis.
  • Evidence: The verify.sh script and measurement.md documentation suggest using go run golang.org/x/perf/cmd/benchstat@latest to compare performance data. This targets an official Go package repository.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the acquisition of performance analysis tools from well-known sources.
  • Evidence: The verify.sh script automatically fetches the benchstat tool from golang.org if it is not already present in the environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 26, 2026, 10:35 AM