This is a cognitive toolkit that gives Claude structured reasoning patterns and memory primitives. Instead of just prompting differently, you get actual tools: thinking models like backtracking and symbolic abstraction, memory cells for state and episodic storage, and multi-agent patterns like debate councils. The Gateway tool routes requests to either direct execution or the Architect, which generates custom agent blueprints by combining these primitives. The Librarian exposes a browsable catalog of all available techniques. Useful when you need more than conversation history, like maintaining state across sessions, enforcing verification steps, or orchestrating multiple reasoning perspectives on complex problems.
mcp-name: io.github.4rgon4ut/sutra
The MCP Context Engineering Engine
Sutra is a Model Context Protocol (MCP) server that transforms how LLMs handle reasoning, memory, and orchestration. It provides a "Standard Library" of cognitive tools (Thinking Models), memory structures (Cells), and multi-agent patterns (Organs).
uv (Recommended)uv tool install context-engineering-mcp
pippip install context-engineering-mcp
Select your agent below and copy-paste the config.
Add to claude_desktop_config.json:
{
"mcpServers": {
"sutra": {
"command": "uv",
"args": ["tool", "run", "context-engineering-mcp"]
}
}
}
Run this in your terminal:
claude mcp add sutra uv tool run context-engineering-mcp
Run aider with the mcp flag:
aider --mcp "uv tool run context-engineering-mcp"
Or add to .aider.conf.yml:
mcp: ["uv tool run context-engineering-mcp"]
Add to ~/.gemini/settings.json:
{
"mcpServers": {
"sutra": {
"command": "uv",
"args": ["tool", "run", "context-engineering-mcp"]
}
}
}
In MCP settings, add a new server:
uv tool run context-engineering-mcpAdd to your config (TOML):
[mcp_servers.sutra]
command = "uv"
args = ["tool", "run", "context-engineering-mcp"]
Sutra automatically analyzes your request to decide the best strategy:
Generates blueprints for custom agents, combining:
understand_question, verify_logic, backtracking, symbolic_abstract.key_value (State), windowed (Short-term), episodic (Long-term).debate_council (Multi-perspective), research_synthesis (Deep Dive).A manual discovery tool (get_technique_guide) that lets you or the agent browse the full catalog of Context Engineering techniques.
Requirements: Python 3.10+, uv (optional but recommended).
git clone https://github.com/4rgon4ut/sutra.git
cd sutra
uv sync --all-extras
# OR
pip install -e ".[dev]"
pytest
MIT
io.github.ericm1018/skillfm-llm-cost-optimizer-openai-anthropic-usage
io.github.mikerawsonnz/llm-orchestration-agent
io.github.mikerawsonnz/authenticated-llm-agent
labforgedev/copilot-memory-mcp
csoai-org/agent-prompt-injection-firewall-mcp
io.github.mikerawsonnz/authenticated-multi-llm-agent