prompt-optimizer
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is strictly advisory. It includes explicit instructions to the agent to avoid executing user tasks, writing code, creating files, or performing any implementation actions, which limits the potential for malicious tool use.
- [SAFE]: Project context detection in Phase 0 involves reading local metadata files (such as
package.json,go.mod, andCLAUDE.md) to identify the project's tech stack. This is a standard practice for providing relevant technical advice, and the skill lacks the network or file-write capabilities required to exfiltrate this information or maintain persistence.
Audit Metadata