Save Money
Security checks across malware telemetry and agentic risk
Overview
This instruction-only skill is mainly a disclosed model-routing helper, but it can automatically hand complex prompts to Sonnet and forward the full task text.
Before installing, make sure you want an automatic cost/quality router that may spawn Sonnet for many complex or uncertain requests. Do not use it with sensitive prompts unless forwarding the full task to the spawned model session is acceptable.
VirusTotal
64/64 vendors flagged this skill as clean.
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.
Complex requests may be handed to a Sonnet session automatically instead of being answered by the current model.
The skill explicitly directs automatic use of a session-spawning tool for complex tasks. This is purpose-aligned with model routing, but users should notice it can change the model used and may incur additional model costs without a separate confirmation step.
If YES → call `sessions_spawn()` IMMEDIATELY. Do NOT attempt to answer yourself.
Use this only if you want automatic model handoff; monitor billing and consider editing the rules if you prefer confirmation before escalation.
If a complex prompt contains private information, that information may be included in the spawned Sonnet session.
The handoff sends the full user task to another model session. That is expected for the skill's routing purpose, but the artifact does not add privacy or minimization guidance for sensitive prompts.
sessions_spawn( message: "<the full task description>", model: "anthropic/claude-sonnet-4-20250514"
Avoid invoking the skill on sensitive content unless you are comfortable with the full prompt being forwarded to the Sonnet session.