openclaw-tool-executor

What This Skill Does

The openclaw-tool-executor serves as the primary bridge between the OpenClaw AI agent and external digital ecosystems. By leveraging Scalekit Connect, this skill standardizes the process of interacting with third-party SaaS platforms and API-driven services. It abstracts away the complexities of authentication, allowing the agent to perform tasks such as querying databases, fetching data from SaaS platforms, managing task lists, and sending communications without requiring manual configuration for every unique service. Whether you need to pull data from a CRM, manage issues in a code repository, or retrieve information from a professional network, this tool manages the secure handshake, tool discovery, and execution pipeline automatically.

Installation

To integrate this skill, ensure your environment is pre-configured with Python 3 and the uv package manager. First, install the necessary authentication helper via the OpenClaw hub: clawhub install openclaw/skills/skills/avinash-kamath/scalekit-agent-auth. Next, configure your .env file with the following variables provided by your Scalekit dashboard: TOOL_CLIENT_ID, TOOL_CLIENT_SECRET, and TOOL_ENV_URL. Optionally, define a TOOL_IDENTIFIER to streamline the default connection selection. Complete the setup by navigating to the skill directory and running uv sync to resolve dependencies, ensuring the executor is ready to handle API requests securely.

Use Cases

This skill is highly versatile, supporting a wide range of operational workflows. You can use it to fetch real-time updates from LinkedIn, automate administrative tasks by reading or writing to Notion documents, sync data across various platforms, or query SQL databases like Snowflake. It is particularly effective for developers needing to automate GitHub issue tracking, marketers requiring instant HubSpot contact lookups, or any user needing to bridge the gap between their AI agent and their productivity stack.

Example Prompts

1. "Search my LinkedIn connections for marketing managers at tech startups and summarize their latest posts."
2. "Look up the status of the 'Q4 Launch' ticket in our GitHub repository and send a Slack message to the engineering team if it's still open."
3. "Fetch the latest customer feedback data from our HubSpot CRM and generate a bulleted summary of recurring issues."

Tips & Limitations

Always ensure that your connection status is set to 'COMPLETED' in the Scalekit dashboard; the executor will ignore draft or incomplete configurations. If multiple connections exist for a single provider, the skill will default to the first discovered, so keep your configuration clean to avoid unexpected behaviors. Be mindful that this tool requires network access and executes code via uv, which may introduce latency depending on the target API's response time.