backend-go-naming
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to interact with and refactor user-provided Go source code, which serves as an untrusted input.
- Ingestion points: The skill utilizes
Read,Glob, andGreptools to ingest content from files in the project workspace (e.g.,.gofiles) into the agent's context as specified inSKILL.md. - Boundary markers: There are no explicit instructions or delimiters defined to ensure the agent ignores instructions embedded within the data it processes (such as malicious comments or string literals within the Go code).
- Capability inventory: The skill is granted
Writepermissions and the ability to execute shell commands viaBash(go:*),Bash(golangci-lint:*), andBash(git:*). This combination of reading untrusted data and possessing write/execute capabilities creates a surface for indirect prompt injection. - Sanitization: No evidence of sanitization, escaping, or validation of the ingested source code is present in the instructions.
Audit Metadata