ml-engineer
PassAudited by VirusTotal on May 3, 2026.
Overview
Type: OpenClaw Skill Name: ah-ml-engineer Version: 1.0.0 The skill bundle consists of a persona definition and operational guidelines for a Machine Learning Engineer agent. The SKILL.md file contains standard industry practices, checklists, and workflows for ML lifecycles without any executable code, suspicious network requests, or malicious prompt injection attempts.
Findings (0)
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.
If connected to real engineering tools, the skill may guide changes to ML pipelines, deployments, monitoring, or retraining workflows.
The skill directs the agent toward production ML implementation and deployment actions. These are consistent with the stated ML engineering purpose, but they could be high-impact if the hosting agent has access to repositories, cloud infrastructure, or deployment systems.
Implementation approach: - Build pipelines - Train models - Optimize performance - Deploy systems - Setup monitoring - Enable retraining
Keep production mutations user-approved, review generated deployment and retraining changes before applying them, and use staging, rollback, and monitoring safeguards.
The agent might overstate model performance or deployment success if it treats the example text as a literal completion message.
The prompt includes a highly specific success message with performance and business metrics. This appears to be illustrative, but it could mislead users if repeated without verifying those results.
Delivery notification: "ML system completed. Deployed model achieving 92.7% accuracy with 43ms inference latency. Automated pipeline processes 10M predictions daily with 99.3% reliability. Implemented drift detection triggering automatic retraining. A/B tests show 18% improvement in business metrics."
Treat the quoted notification as an example only, and require the agent to report only metrics that were actually measured or supplied by the user.