The Agent Skills Directory

[SAFE]: The skill serves as a security-focused persona and knowledge base for auditing and developing Go applications. It covers critical security domains including injection prevention, cryptographic best practices, and filesystem safety.
[SAFE]: Integration with external tools involves only well-known and authoritative security scanners (such as gosec and govulncheck) from trusted official repositories.
[SAFE]: The provided evaluation dataset (evals.json) correctly asserts security best practices, ensuring that the agent's performance is measured against high-security standards for scenarios like SQL injection prevention and secure password hashing.
[INDIRECT_PROMPT_INJECTION]: The skill's architecture involves ingesting and analyzing external source code which constitutes an attack surface.
Ingestion points: The agent reads project files using Read, Glob, and Grep tools.
Boundary markers: The instructions provide clear guidance on identifying trust boundaries and implementing defense-in-depth layers to isolate untrusted data.
Capability inventory: The skill uses Write and Edit for code remediation and Bash for running Go-specific security linters.
Sanitization: The skill contains logic to teach and verify sanitization in target codebases; it does not process external data in an executable context that would compromise the agent itself.

backend-go-security