proofline-onboarding
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a configuration for an onboarding assistant. It explicitly instructs the agent to avoid common security pitfalls:
- Credential Safety: The instructions strictly forbid requesting or accepting credentials from the user and state the session has no access to sensitive secrets.
- Least Privilege: The agent is restricted to reading specific non-sensitive file types (Terraform, YAML manifests, documentation) and cannot execute arbitrary shell commands or perform network operations outside of the provided tools.
- Human-in-the-Loop: All critical operations, such as cloud connections, access approvals, and final configuration confirmations, are deferred to the Proofline console or require explicit user action via the CLI.
- Indirect Injection Mitigation: The skill explicitly warns that repository content and guidance files (like
AGENTS.md) must be treated as untrusted data, not instructions to the AI agent.
Audit Metadata