Model Healthcheck
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only skill coherently tests configured models, but users should expect it to read model configuration and make concurrent test calls to configured providers.
This skill appears safe and purpose-aligned. Before running it, be aware that it will read the configured model list and may send a small test prompt to every configured model provider concurrently, which could use quota or surface provider/authentication errors.
Findings (2)
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.
Running the skill may contact every configured model provider at once and could consume quota or produce provider errors.
The skill directs the agent to create multiple model test sessions concurrently. This is central to the healthcheck purpose, but it can result in multiple provider calls, costs, or rate-limit effects.
For each model, use `sessions_spawn` to run a test ... Fire all tests concurrently (multiple `sessions_spawn` calls at once), do not test sequentially.
Use it when you intend to test all configured models; specify a single model if you want to limit calls.
The resulting report may disclose which model providers and model IDs are configured in the environment.
The skill reads gateway configuration to identify configured providers and models. This access is expected for its purpose, and the instruction narrows use to provider/model combinations rather than credentials.
Use `gateway config.get` to read the current config and extract all `provider/model` combinations from `models.providers`.
Only run it in contexts where sharing configured provider/model names with the current user is acceptable.