Query Router

v1.1.0

Unified query router combining content-type detection, complexity scoring, and prefix-based routing. Routes queries to the optimal model with safety features...

0· 50·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name/description (route queries to optimal models) match the code and SKILL.md: prefix/content-type/complexity routing, mapping to cloud/local models. The use of the 'openclaw' CLI and a localhost model tags endpoint is coherent for a router that needs current model state and available models.
Instruction Scope
Runtime instructions and the code operate within the claimed domain: classify input, decide model, optionally call 'openclaw' to check/verify status, and write local audit logs. The skill references only local files (.locks, .logs) and localhost HTTP; it does not exfiltrate to external servers or request unrelated system secrets.
Install Mechanism
No install spec is provided (instruction-only). The package includes Python scripts but does not download or extract remote artifacts. This is the lowest-risk install pattern for a script-based skill.
Credentials
The skill declares no required env vars or credentials and the code does not read secrets. It invokes a local CLI ('openclaw') and contacts http://localhost:11434 which are proportional to discovering models; no unrelated service tokens are requested.
Persistence & Privilege
always is false and the skill does not attempt to modify global skill config. It creates local lock and log files under scripts/.locks and scripts/.logs, which is expected for routing/audit behavior and does not indicate elevated privileges.
Assessment
This skill appears coherent for routing queries: it classifies queries, consults a local model endpoint (http://localhost:11434/api/tags) and the 'openclaw' CLI, and writes audit logs under scripts/.logs and lock files under scripts/.locks. Before installing, verify you trust the skill source (homepage unknown) and that running 'openclaw' and a local model registry on localhost is expected in your environment. Review the included Python files yourself or run the scripts in a safe/test workspace; note they log up to the first 100 characters of queries to a JSONL file, so avoid sending sensitive secrets through queries if you enable this skill.

Like a lobster shell, security has layers — review code before you run it.

latestvk9711hk23wvbn04aq8b1p5xdfh843xbk

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments