A Model Context Protocol (MCP) server that integrates with n8n, providing tools for workflow and execution management via the n8n API.
3.0 KiB
3.0 KiB
Instructions for Using CRCT v7.0
These instructions assume you’re using the Cline VS Code extension with CRCT v7.0. The system automates most tasks via the LLM, but some manual steps are needed to get started.
Prerequisites
- VS Code: Installed with the Cline extension.
- Python: 3.8+ with
pip. - Git: To clone the repo.
Step 1: Setup
-
Clone the Repository:
git clone https://github.com/RPG-fan/Cline-Recursive-Chain-of-Thought-System-CRCT-.git cd Cline-Recursive-Chain-of-Thought-System-CRCT- -
Install Dependencies:
pip install -r requirements.txtIncludes
sentence-transformersfor embeddings. -
Open in VS Code:
- Launch VS Code and open the
cline/folder.
- Launch VS Code and open the
-
Configure Cline:
- Open the Cline extension settings.
- Paste the contents of
cline_docs/prompts/core_prompt(put this in Custom Instructions).mdinto the system prompt field.
Step 2: Initialize the System
-
Start the LLM:
- In the Cline input, type
Start.and run it. - The LLM will:
- Read
.clinerules(or create it if missing). - Load the
Set-up/Maintenanceplugin. - Initialize core files in
cline_docs/.
- Read
- In the Cline input, type
-
Follow Prompts:
- The LLM may ask for input (e.g., project goals for
projectbrief.md). - Provide concise answers to help it populate files.
- The LLM may ask for input (e.g., project goals for
-
Verify Setup:
- Check
cline_docs/for new files (e.g.,dependency_tracker.md). - Ensure
[CODE_ROOT_DIRECTORIES]in.clineruleslistssrc/(edit manually if needed).
- Check
Step 3: Populate Dependency Trackers
-
Run Initial Setup:
- Input:
Perform initial setup and populate dependency trackers. - The LLM will:
- Identify code roots (e.g.,
src/). - Generate
dependency_tracker.mdanddoc_tracker.mdusingdependency_processor.py. - Suggest and validate dependencies.
- Identify code roots (e.g.,
- Input:
-
Manual Validation (if prompted):
- The LLM may present dependency suggestions (e.g., JSON output).
- Confirm or adjust characters (
<,>,x, etc.) as prompted.
Step 4: Plan and Execute
-
Enter Strategy Phase:
- Once trackers are populated, the LLM will transition to
Strategy(check.clinerules). - Input:
Plan the next steps for my project. - Output: New instruction files in
strategy_tasks/orsrc/.
- Once trackers are populated, the LLM will transition to
-
Execute Tasks:
- Input:
Execute the planned tasks. - The LLM will follow instruction files, update files, and apply the MUP.
- Input:
Tips
- Monitor
activeContext.md: Tracks current state and priorities. - Check
.clinerules: Shows the current phase and next action. - Debugging: If stuck, try
Review the current state and suggest next steps.
Notes
- CRCT v7.0 is a work-in-progress. Expect minor bugs (e.g., tracker initialization).
- The LLM handles most commands (
generate-keys,set_char, etc.) automatically. - For custom projects, populate
src/anddocs/before starting.
Questions? Open an issue on GitHub!