SShortSingh.
Back to feed

RepoJanitor: AI Tool Automates Routine Repository Cleanup for Developers

0
·2 views

RepoJanitor is an open-source AI maintenance tool designed to handle the repository upkeep developers often postpone, such as removing duplicate files, stale comments, and outdated documentation. It scans a GitHub repository or local codebase using deterministic analysis, then applies an open-weight AI model to reason about the findings. A deterministic safety layer sits between the AI and the filesystem, and each proposed action is labeled SAFE, REVIEW, or BLOCKED. Changes are made on an isolated branch rather than main, and developers keep final control over what gets applied. The project was built as an entry for the Hacktoberfest Open-Source AI Challenge.

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
ProgrammingHacker News ·

Oregon Police Quote $2.5M Fee to a Local Outlet for Records on One Officer

A local news outlet in Oregon has been quoted roughly $2.5 million by police for public records concerning a single officer's misconduct. The Intercept reports that the high figure has drawn scrutiny over how agencies price public records requests. Critics argue such charges can effectively block journalists and the public from accessing information about government employees. The story frames the fee as a potential barrier to transparency in police accountability. The article was shared on Hacker News, where it drew modest discussion.

0
ProgrammingDEV Community ·

Installing Self-Hosted Jitsi Meet on Ubuntu 24.04 for Private Video Calls

A guide explains how to deploy Jitsi Meet, an open-source video conferencing platform, on a self-managed Ubuntu 24.04 server. It highlights advantages over services like Zoom and Google Meet, including no guest accounts and no time limits. The setup covers hostname and DNS configuration, adding the Prosody and Jitsi package repositories, and firewall rules for web and UDP media traffic. It also walks through installing jitsi-meet with a Let's Encrypt certificate, which is needed for mobile app compatibility. Recommended hardware is at least 2 GB of RAM, with 4 GB advised for real meetings.

0
ProgrammingDEV Community ·

PHP Team Warns of Unpatched RCE Flaw, Issues Interim Configuration Mitigations

The PHP security team has warned of a newly disclosed remote code execution vulnerability that does not yet have an official patch. The flaw affects only certain non-default configurations, allowing an attacker to run arbitrary code on the server via a crafted request. In the absence of a fix, the team has published interim mitigations, including disabling risky functions and remote URL inclusion and restricting filesystem access. The alert comes amid other recent incidents, such as a compromised LiteLLM package distributing info-stealer malware and a Railway breach prompting secret rotation. Engineers are advised to audit non-default PHP settings, apply the recommended changes, and monitor for the forthcoming official patch.

0
ProgrammingDEV Community ·

Daylight: A Local AI Planner That Schedules Work and Personal Tasks Together

Daylight is a personal assistant that turns a task list into a proposed day, mixing work and personal items while accounting for priorities, time estimates, fixed appointments and breaks. Users set available hours, enter tasks as lines or through a form, and a local open-weight model suggests categories, priorities and durations that the user can review before scheduling. Tasks that do not fit are listed under a "For another time" section with an explanation, and the day can be replanned when plans change. The project is open source under the MIT license, built with plain HTML, CSS and JavaScript, and runs locally with a small Node.js server. Early browser testing led the developer to restrict the model to a simpler output format, since it had invented times and dropped tasks, and the results were informal rather than a formal accuracy evaluation.