A specialized utility to prevent OpenClaw from stripping caching headers for third-party OpenAI-compatible API providers.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install responses-third-party-prompt-cache-patch
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 responses-third-party-prompt-cache-patch 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 Responses Third Party Prompt Cache Patch?
The Responses Third Party Prompt Cache Patch is a technical utility designed for the Openclaw Skills ecosystem. It addresses a specific limitation where the agent strips prompt_cache_key and prompt_cache_retention metadata from third-party API responses. By applying a surgical patch to the OpenClaw distribution bundle, developers can ensure that advanced caching features of alternative LLM providers remain intact.
This tool is essential for users who leverage non-OpenAI endpoints that support these specific caching parameters to reduce latency and costs. It operates as a local patch, providing a workflow that includes dry-runs, automatic backups, and syntax validation to ensure the stability of the agent installation.
Responses Third Party Prompt Cache Patch Use Cases
- Enabling prompt caching on third-party OpenAI-compatible endpoints to improve performance.
- Reducing API latency by preserving cache keys in the agent response handling logic.
- Maintaining custom metadata for cost-tracking and efficiency within Openclaw Skills environments.
- Local patching of OpenClaw installations without waiting for upstream refactors or configuration changes.
How Responses Third Party Prompt Cache Patch Works
- The script locates the OpenClaw installation root by resolving the openclaw executable path or using a provided root flag.
- It scans the dist directory for the appropriate JavaScript bundle, prioritizing embedded files.
- A timestamped backup of the original bundle is created before any modifications are written.
- The patcher applies a narrow modification to the shouldStripResponsesPromptCache function to allow specific metadata retention.
- The modified file is validated using node --check to ensure no syntax errors were introduced.
- Upon successful validation, the user restarts the gateway to activate the changes across their Openclaw Skills.
Responses Third Party Prompt Cache Patch Setup
To install this skill into your workspace, use the following command:
clawhub install responses-third-party-prompt-cache-patch --workdir ~/.openclaw/workspace
Once installed, navigate to the skill directory and execute the patcher:
python3 scripts/patch_prompt_cache.py --dry-run
python3 scripts/patch_prompt_cache.py
openclaw gateway restart
Responses Third Party Prompt Cache Patch Data Schema & Taxonomy
| Component | Description |
|---|---|
| scripts/patch_prompt_cache.py | Primary automation script for applying the cache retention patch. |
| scripts/revert_prompt_cache.py | Safety utility to restore the system to its pre-patched state. |
| Backups | Local timestamped copies of the dist bundles stored for recovery. |
| Target Logic | Specific modification of the shouldStripResponsesPromptCache(model) function. |
Responses Third Party Prompt Cache Patch Advanced Features
- Dry-run capability to preview file changes without modifying the actual installation.
- Integrated rollback system that automatically identifies and restores the latest matching backup.
- Syntax validation using Node.js to prevent gateway crashes after patching.
- Upgrade-aware reapplication logic that detects if a bundle needs a fresh patch after an agent update.
- Support for custom root paths to test patches on fixtures or alternative installations.
SKILL.md
Loading
Related Openclaw Skills
Accounts Payable Agent
An autonomous AI agent designed to manage vendor payments, contractor invoices, and recurring bills across crypto and fiat channels with strict audit integrity.
CNC Intelligent Quoting System (RAG-based)
A Retrieval-Augmented Generation (RAG) system designed to automate manufacturing cost estimation and CNC part quoting with high precision.
Feishu Meeting Minutes Generator
An automated tool to convert Feishu/Lark meeting transcripts into structured Markdown minutes and professional PDF documents.
Agentic Identity Trust
A robust cryptographic framework for autonomous AI agents to establish verifiable identity, scoped authorization, and immutable audit trails.
Academic Study Planner
A professional learning consultant specialized in creating personalized, evidence-based study plans for major Chinese academic and professional examinations.
1inch DEX Aggregator
A powerful crypto trading skill that utilizes 1inch v5.2 API to find and execute the most cost-effective token swaps across multiple decentralized exchanges.
Featured*
