SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World
Back to feed
Programming

CodeClone 2.1 aims to give AI coding agents auditable change boundaries

0
·9 views

A developer has released CodeClone 2.1.0a1, an open alpha tool designed to add structural change control to AI-assisted Python development. The tool addresses a core problem with AI agents: a passing diff and green tests do not reveal whether the agent stayed within its intended scope or silently modified unrelated code. Before any edit, the agent declares its intent and target file scope, and CodeClone responds with a defined blast radius, explicit do-not-touch boundaries, and an edit_allowed flag. After the edit, the tool reconciles what was declared against what actually changed, producing an auditable record called a Patch Trail. This allows reviewers to evaluate an agent's work against a stated contract rather than reverse-engineering intent from the diff alone.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Developer Open-Sources Claude Basecamp, a Unified Dashboard for Claude Code Workflows

A developer has released Claude Basecamp, an open-source tool designed to serve as a central operating system for working with Anthropic's Claude Code across multiple repositories. The project, launched publicly this week, offers features such as automated test-status checks, session recovery after crashes, and a 'reflexes' system that learns from past user corrections to avoid repeating mistakes. Built in roughly 4,000 lines of plain Node.js, it requires no installation and runs locally at localhost:4747 with no telemetry or external data transmission. The creator acknowledges the project began as a personal fix for keeping tests green but has grown into a broader vision for managing every Claude Code session, goal, and workflow in one place. Contributions are being actively sought, with the codebase released under the permissive Unlicense, and the repository is available on GitHub.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Sparsi framework cuts AI agent token costs and latency using deterministic DAG tools

Developers building ReAct-based AI agents in production often face high token costs, slow latency, and occasional hallucinations due to LLMs re-deriving the same logic on every request. To address this, a team built Sparsi, a framework that offloads predictable sub-routines from the LLM prompt into deterministic Directed Acyclic Graph (DAG) structures called Macro-Tools. These DAGs can be exposed to existing agents as single Model Context Protocol tools, letting the agent delegate reliable, testable logic without repeated reasoning overhead. In a benchmark test on a five-step customer support triage task using 100 samples, Sparsi achieved 100% pipeline accuracy at 1.73 seconds average latency consuming 61,158 tokens, compared to 94% accuracy, 3.06 seconds, and 258,050 tokens for a pure LangChain ReAct agent. Sparsi is available as both a standalone pipeline solution and as a plug-in tool builder for existing agent architectures, with a pre-built operator library included.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Schola Interlingua: Free Open-Source App Launches for Learning Interlingua IALA

Schola Interlingua is a free, open-source platform designed to help users learn Interlingua IALA, a constructed language known for its transparent vocabulary and morphology. The platform combines lessons, readings, review tools, and progress tracking into a structured educational path. Developed by Ian Blas, a Buenos Aires-based software developer, the project's source code is publicly available on GitHub for collaboration and review. The app is accessible across multiple devices and can be downloaded via the App Store and Google Play.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Open-Sources Claude Basecamp, a Reconciliation Tool for Claude Code Users

A developer has open-sourced Claude Basecamp, a local management tool designed to bring Kubernetes-style reconciliation logic to AI-assisted coding workflows using Claude Code. The tool continuously checks that user-defined conditions remain true across one or more repositories, automatically dispatching fix runs when drift is detected. It also includes a 'Reflexes' feature that mines past Claude Code session transcripts to block repeated mistakes machine-wide, and a Session Rescue function that can resume interrupted Claude Code tasks mid-context. Built in roughly 4,000 lines of plain Node.js and vanilla JavaScript, it runs entirely locally with no telemetry and is released under the public-domain Unlicense. The developer is inviting contributors to help extend the project, particularly around cross-repo token budgeting, cost guardrails, and additional routine templates.

0 comments Read more at DEV Community