UACOS (by developer caotiensinh) is a local-first code intelligence, context-compression, and release-gate toolkit designed to operate as an agent harness and orchestration framework. It acts as an execution and guardrail layer for autonomous coding tasks, bridging developer environments with LLMs while enforcing strict boundaries on code modifications.
Context Compression & Project Memory — Maintains typed project memory to optimize prompt token usage and reduce LLM overhead across coding sessions. - Patch-Scope Safety Gates — Enforces boundary checks on agent-generated edits to prevent unwanted modifications beyond defined task scopes. - Transaction Rollback — Provides local rollback mechanisms to revert failed or destructive code changes executed during automated tasks. - Built-in MCP Server — Exposes code intelligence and orchestration tools directly to external agents via Model Context Protocol. - Secret Scanning & Security Checks — Scans working directories and diffs automatically to block secret leakage before applying patches. - Fullstack JS/TS Impact Analysis — Evaluates code changes across JavaScript and TypeScript dependencies using function, path, and API endpoint heuristics. - LLM Cache & Budget Controls — Monitors token spend and caches model responses locally to minimize API costs during agent loops.
Use cases
Use Case: Safe Autonomous Code Refactoring Developers can run automated tasks under Autopilot Mode while patch-scope gates and transaction rollback ensure code modifications remain contained within targeted files and can be safely reverted if tests fail.
Use Case: MCP-Integrated Agent Assistance Engineers using MCP-compatible clients (such as Claude Desktop or IDE extensions) can connect directly to UACOS's built-in MCP server to query local project memory, run impact rankings, and fetch compressed code context.
Use Case: Pre-Commit Release Gating & Secret Verification Teams can integrate UACOS into local pipelines to scan modified code for hardcoded secrets, test integrity, and cross-file JavaScript/TypeScript impact before committing changes.
Who it is for
UACOS is designed for software engineers and autonomous agent builders who need guardrails, rollback protection, and context compression when running coding models against local codebases. It is less suited for non-technical users seeking a simple prompt-and-chat interface without CLI or MCP workflow configuration.
Install guide
Installation Clone the repository and install the toolkit locally (standard Python environment setup): bash git clone https://github.com/caotiensinh/uacos.git cd uacos pip install -e .
First Run Initialize configuration and explore the available CLI surface: bash # Inspect available commands uacos --help # Initialize VS Code settings or start the local dashboard if desired uacos vscode-init
Example Session Start the built-in MCP server or run an impact analysis: bash # Start the MCP server for external agent integration uacos mcp serve # Run safety and release checks across the workspace uacos release-gate --check
Continue — Continue is an open-source AI code assistant and agent framework originally developed by the Continue team (continuedev) and recently acquired…
Mneme — Mneme is an open-source orchestration harness and memory governance framework developed by MnemeHQ designed to prevent architectural drift across the…
wt — wt is an open-source CLI harness and workspace orchestration tool maintained by Sahil Ahuja on GitHub. It is built to…
Grite — Grite is an open-source terminal coding agent harness developed by Neul Labs. Implemented in Rust, it provides a lightweight command-line…
Agent Memory System — Agent Memory System is an open-source agent orchestration harness developed by RavByte-AI. It provides structured memory retention and retrieval mechanisms…
AgentPack — AgentPack is an open-source orchestration harness and configuration management framework developed by Vishal (vishal2612200) for multi-agent coding environments. It standardizes…