A specialized diagnostic toolkit for validating LLM API protocol compatibility, first-token detection, and performance drift for coding agents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install api-quality-check
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install api-quality-check using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is API Quality Check?
API Quality Check is a script-first utility designed to ensure that AI coding models and their respective API endpoints meet the technical requirements for advanced agentic workflows. It specifically focuses on verifying first-token detection, logprob tracking, and baseline-vs-current drift checks using LT-lite and B3IT-lite methodologies. By incorporating these Openclaw Skills into your pipeline, you can programmatically determine if an OpenAI, Anthropic, or OpenAI-Compatible provider is suitable for headless coding tasks.
The tool addresses the common problem of model volatility and hidden protocol mismatches. It allows developers to smoke-test endpoints like DeepSeek, SiliconFlow, Kimi, and Ark to verify capability fit before deploying agents that rely on precise token-level control or specific header configurations.
API Quality Check Use Cases
- Validating if a new LLM provider supports logprobs required for LT-lite detection.
- Running daily automated smoke tests to detect performance drift in production coding models.
- Normalizing varied API provider configurations into a standard JSON format for agent portability.
- Testing first-token response latency and capability across multiple OpenAI-compatible endpoints.
How API Quality Check Works
- Initialize a provider configuration file that normalizes the API root and authentication headers.
- Execute a capability smoke test to identify if the endpoint supports logprobs or first-token text at minimum token limits.
- Based on smoke test results, the system recommends a specific detector (LT-lite or B3IT-lite).
- Generate a performance baseline for the model to capture its current behavior profile.
- Run periodic detection tasks to compare the live API performance against the saved baseline to identify drift.
- Export results as structured JSON for CLI consumption or human-readable HTML for visual reporting.
API Quality Check Setup
Set the environment variables for the core scripts within your Openclaw Skills directory:
export CODEX_HOME="${CODEX_HOME:-$HOME/.codex}"
export APIQ="$CODEX_HOME/skills/api-quality-check/scripts/api_quality_check.py"
Initialize a configuration for a specific provider:
python "$APIQ" init-config \
--provider "OpenAI-Compatible" \
--base-url "https://api.your-provider.com/v1" \
--api-key "$API_KEY" \
--model-id "model-name" \
--config-output ./provider.json
Run a smoke test to verify capabilities:
python "$APIQ" smoke --config ./provider.json --html-output ./smoke.html
API Quality Check Data Schema & Taxonomy
The skill manages data using a file-based approach to ensure compatibility with terminal agents and Openclaw Skills workflows:
| Data Component | Format | Description |
|---|---|---|
| Provider Config | JSON | Stores base URL, model ID, and custom headers like User-Agent or Extra-Body. |
| Capability Report | JSON/HTML | Result of smoke tests indicating lt_supported and b3it_supported flags. |
| Baseline Manifest | JSON | A snapshot of model behavior used for future drift comparisons. |
| Detection Report | JSON | Detailed metrics on stability, token counts, and drift probability. |
API Quality Check Advanced Features
- Multi-Provider Batching: Run capability checks across an entire list of providers using a single shell command.
- Auto-Reasoning Mitigation: Automatically detects and disables thinking/reasoning blocks for providers like GLM or Zhipu to ensure clean output.
- Protocol Normalization: Intelligent URL handling that converts full completion paths back to API roots automatically.
- Scheduled Drift Monitoring: Includes a dedicated daily wrapper script for automated health checks and date-based report archiving.
- Custom Header Support: Pass complex JSON headers for specialized endpoints like Kimi's coding-specific API paths.
SKILL.md
Loading
Related Openclaw Skills
DOCX Document Master
A professional-grade toolkit for programmatically creating, editing, and analyzing Microsoft Word documents with high-fidelity formatting.
App Ads Skill
A comprehensive guide for executing high-performing mobile app install campaigns and user acquisition strategies across major ad platforms.
Claude API & Agent SDK Integration
A comprehensive development framework for building LLM-powered applications using the Anthropic Claude API and the specialized Agent SDK.
Find Skills
A specialized discovery tool for listing and searching through all installed AI agent capabilities within a workspace.
Invoice Chaser
A professional billing automation tool that generates escalating payment reminder emails tailored to specific business verticals and days past due.
WeryAI Video Subtitle Erase
A specialized AI tool designed to identify and erase burnt-in subtitles or on-screen text from public video URLs via the WeryAI API.
Featured*
