backend-go-concurrency
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves reviewing and auditing external source code, creating an attack surface where malicious instructions in that code could influence the agent's behavior.
- Ingestion points: Code review and codebase audit modes ingest untrusted source code and PR diffs via the
ReadandGreptools. - Boundary markers: The instructions do not specify the use of delimiters or 'ignore' directives for content within the audited files.
- Capability inventory: The skill allows the agent to edit files and execute Go-specific shell commands through the
Edit,Write, andBash(go:*)tools. - Sanitization: There are no provisions for sanitizing or escaping the content of processed code files before analysis.
- [EXTERNAL_DOWNLOADS]: The skill instructions and code examples reference external libraries to facilitate leak detection and concurrency management.
- Evidence: The skill references
go.uber.org/goleakand sub-packages fromgolang.org/x/sync. These are official or highly reputable extensions within the Go ecosystem.
Audit Metadata