AI Task Router Auto-Route Prompts by Complexity (Local/Cloud)

Security checks across static analysis, malware telemetry, and agentic risk

Overview

This appears to be a simple in-memory prompt-routing helper, with no evidence of hidden network access, credential use, persistence, or destructive behavior.

This skill looks benign and narrowly scoped. Before installing, check whether your OpenClaw setup will actually dispatch routed prompts to cloud providers such as Sonnet or Opus, and review any custom routing rules so sensitive prompts or expensive tasks are not routed unexpectedly.

Static analysis

No static analysis findings were reported for this release.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

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.

What this means

Some prompts may be routed to cloud or premium models depending on configuration, which can affect privacy and cost.

Why it was flagged

The skill is intended to select between local and cloud-capable model tiers. Although the provided code only returns a model name and does not transmit data, an integration using these decisions could send prompt content to external providers.

Skill content
| Simple | ... | Local (Ollama) | ... | Moderate | ... | Mid-tier (Sonnet) | ... | Complex | ... | Premium (Opus) |
Recommendation

Confirm which models the host agent will actually call, and require explicit confirmation for sensitive or high-cost prompts if needed.

What this means

Bad custom rules could cause unexpected local/cloud model choices or disrupt routing.

Why it was flagged

Custom regex rules can change routing behavior. This is purpose-aligned customization, but unreviewed or overly broad rules could route prompts to the wrong model tier or introduce reliability issues.

Skill content
if (config.customRules) { ... this.addRule(rule.name, new RegExp(rule.pattern, rule.flags || 'i'), rule.model, rule.reason); }
Recommendation

Only load custom rules from trusted configuration, review model targets, and test rules before using them for automatic routing.

What this means

It may be harder to verify the publisher or track upstream changes.

Why it was flagged

The registry metadata does not identify a canonical source or homepage. The included code is small and has no install-time dependencies, so this is a provenance note rather than a security concern.

Skill content
Source: unknown; Homepage: none
Recommendation

Review the included source before installing and prefer pinned, trusted versions if deploying broadly.