Prompt Refiner
Security checks across static analysis, malware telemetry, and agentic risk
Overview
This is a coherent prompt-formatting helper with simple local CLI code and no evidenced credential, network, persistence, or destructive behavior, though users should review prompts before letting an agent act on them.
This skill appears safe to install as a prompt-refinement helper. Before using it for real work, make sure the refined prompt names the correct account, file, service, or destination, and do not let the agent perform state-changing actions without your explicit approval.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
An agent may proceed after refining a request if it considers the task safe or obvious.
The skill is mainly for prompt refinement, but it also tells the agent when to continue into execution. It includes a confirmation requirement for destructive or complex actions, so this is a usage-scope note rather than a concern.
- Destructive/complex actions: Show 1-sentence summary → get confirmation - Safe/obvious tasks: Execute directly
Use the skill to produce clearer prompts, but require explicit confirmation before any account action, external post, service restart, deployment, deletion, or other state-changing operation.
Users may trust the refined prompt more than warranted and skip reviewing important details.
The README makes strong performance and reliability claims that are not substantiated by the artifacts. This could encourage overconfidence in generated prompts.
AI executes perfectly on the first try ... Improves AI accuracy by 40-60% (no more hallucinations from unclear input)
Treat the reliability claims as promotional; review the generated prompt and verify targets, accounts, paths, and intended actions before execution.