Developer Corner

A catalog where AI agents file technical discoveries, failed approaches, evidence, and next steps for other agents to unlock and continue.

An iPhone and iPad SSH terminal that verifies server host keys before authentication, with a $4.99 one-time price.

An open measurement platform for testing whether agent skills work, what they cost, and which models they suit.

A local-first board for directing coding agents, reviewing their work, and merging completed changes.

EffectProbe audits agent tool calls by marking arguments, observing where the values surface, and compiling the results into an enforceable contract.

Engwire detects meaningful GitHub and deployment changes, explains them for different audiences, and sends concise updates to Slack.

A cryptographically signed registry posts concrete coding problems and lets other people submit solutions for independent verification.

GQueries presents persistent memory, authorized evidence, and fail-closed controls for production AI systems.

A Boston security engineer's portfolio covering MCP server audits, supply-chain vulnerabilities, and open-source scanning tools.

A native macOS desktop pet that lets an existing AI coding agent run commands, edit files, and ask for permission out loud.

A browser-based utility that tests a file path against pasted .gitignore contents and explains Git’s verdict.

A browser-based checker that compares agent-facing Markdown instructions with scripts, paths, versions, lockfiles, and other repository evidence.

A Raspberry Pi Pico-based device and free Windows app for extracting Super Famicom ROMs over USB.

A measured comparison and buying advisor for running language models on local hardware.

An open-source tool that finds, validates, and ranks logo assets exposed by a website.

A self-hosted MCP memory server that gives multiple AI clients one shared store, filtered by policies you define.

A self-hosted Certificate Transparency monitor that watches for unexpected certificates and lookalike domains before they become phishing infrastructure.

A free macOS mock API server for scripting repeatable failures, retries, timeouts, and other awkward HTTP edge cases.

A pre-release .NET package gives Minimal APIs an endpoint-class structure while keeping ASP.NET Core's routing, authorization, OpenAPI, and other native machinery underneath.

An open-source workspace and research catalogue for testing how AI changes education, assessment, and evidence of learning.

A readable observatory for 1F916, a public forum populated by AI agents, with maps, citizen profiles, network statistics, and cryptographically chained snapshots.

An MIT-licensed, offline deck-generation tool that turns an AI’s instructions into editable PowerPoint files with real text, shapes, charts, and layouts.

An open-source, self-hosted workspace that turns STL, 3MF, OBJ, and PLY folders into a searchable 3D model library.

A local proxy replaces names, addresses, phone numbers, organisations and bank details with stand-ins before sending requests to an AI service, then restores them in the response.

Riot is an open-source desktop AI agent that reads projects, edits files, runs commands, browses, and produces documents locally.

A simulator-verified framework for turning robot interaction into inspectable, reusable code-backed skills.

A local orchestrator directs coding-agent CLIs through security research, evidence review, and guarded Git acceptance.

A tiny POSIX shell that connects to a model, runs tools, and writes code in the directory where it is launched.

A no-signup collection of developer utilities for SEO, email security, DNS, networking, app links, CORS, regex, Caddy, and digital wallets.

tentkotta gathers six open-source projects, including offline Android apps, browser tools, Python utilities, and a vault-migration script.

An unaffiliated parody community project that reads two local AI-coding-tool usage numbers and anonymously adds developers to a public token-burn leaderboard.

A command-line diagnostic tool that reads ROS 2 logs and related recordings, then turns failures into causal reports tied to files and lines in a code workspace.

The Unified Harness Protocol is an open HTTP standard for starting, tracking, continuing, cancelling, and retrieving results from complete agent runtimes.

An Athena-first desktop app streams Muse EEG live, displays contact quality and bands, and records sessions to CSV or EDF.

A persistent Linux computer that agents can drive through an API while a person watches or takes the keyboard.

A searchable index of Australian APIs and catalogues, with publisher details, authentication notes, live-status checks, and connection recipes.

A free desktop app that keeps Claude Code agents, terminals, files, Git, notes, tasks, and machine monitoring together in one project window.

A private browser playground that runs HTML, Markdown, JSX, TSX, SVG, XML, CSS, and JSON with live preview, console output, and optional direct-to-provider AI editing.

A public AI negotiation game and benchmark where agents compete in ten-minute duels for virtual ContextCoins.

A local-first Markdown workspace that connects a notes folder with a read-only code repository and exposes the combined graph to AI agents over MCP.

A read-only monitor that tracks changes to a Slack workspace’s security posture over time.

An infrastructure layer that lets products run Codex, Claude Code, Hermes, and other agent harnesses with sessions, streaming, files, and artifacts.

HarnessRouter runs Codex, Claude Code, Hermes, and other agent harnesses in isolated sandboxes so products can receive code, files, documents, and other finished artifacts.

一个中文创作者社区，精选、实测并整理 GitHub 上的 AI、自动化、编程和内容创作项目。

A virtual-server provider offering five hourly billed VPS plans, from €4 a month, with root access and no listed paid extras.

A Zig-built terminal AI agent harness that keeps structured sessions alive in a background daemon while clients come and go.

An open-source tool that turns requirements from documents, images, URLs, or plain text into BDD acceptance criteria, coding constraints, test scaffolds, and verification reports.

A fund-accounting platform puts the accounting book and investment book on one ledger, with continuous checks across complex fund structures.

A non-custodial Bitcoin mining pool prototype that uses shared work snapshots instead of a pool wallet, coordinator, or parallel blockchain.

A macOS and Windows beta recorder that follows work across approved displays and applications while masking protected apps.

Loreqo turns project decisions, conventions, and constraints into shared context for Claude Code, Cursor, Codex, Gemini, and other coding agents.

An open-source WordPress plugin that stores every language version in the original post instead of creating duplicates.

A browser-based tool that compares Jira automation JSON exports and checks them for risky rules, unsafe actions, loops, missing guards, and fan-out.

An open temporal data format and reference engine for representing uncertain dates, intervals, calendar periods, and media offsets.

An open standard for turning data meaning, computations, assumptions, and expert know-how into portable executable Recipes.

An Algerian self-hosted AI secure web gateway for allowing, blocking, and inspecting web traffic with verdicts and logs.

A curated index of Australian public-data APIs and catalogues, organized by jurisdiction, publisher, authentication method, live status, and connection instructions.

A design preview for a structured whiteboard connecting documents, live data, verified artifacts, and AI review.

BytePort ingests master video files, creates adaptive renditions, and delivers player-ready HLS or DASH output.

An iOS protocol client for remotely controlling a self-hosted DeepSeek Harness host, including sessions, approvals, questions, model settings, and notifications.

EvioNav presents an event-camera navigation layer for drones and autonomous vehicles operating without reliable satellite positioning.

A short-link service with QR codes, click analytics, custom domains, a REST API, and an MCP server for AI agents.

An open-source local SMTP relay designed to let printers, scanners, and legacy applications send mail through Microsoft 365 with OAuth and TLS.

HP is a v0.1.0 systems programming language whose keywords can be written in either Chinese or English.

Hitch is a Ruby gem that lets Claude, ChatGPT, and Cursor call registered Rails tools as a signed-in user.

A free collection of browser-based tools for formatting, validating, converting, comparing, encoding, and generating developer data.

A visual ETL builder that turns drag-and-drop data pipelines into editable SQL.

A research-preview pipeline turns natural-language geometry problems into editable GeoIR and Lean-checked formal proofs.

A micropayment layer for publishers and readers, handling charges as small as €0.001 through shared or site-specific wallets.

A browser-based tool inspects an open page and packages its visual evidence—tokens, typography, spacing, components, assets, and states—for AI-assisted building.

A daily guessing game that asks you to identify a hidden AWS service before six requests run out.

A push-notification interface that lets AI agents ask people questions and receive answers from a phone.

A Python package and planned inference platform for exploring reionization observables without rerunning expensive physics simulations.

A support-automation evaluator that tests proposed AI handling against held-out synthetic tickets.

A public catalog for discovering, comparing, quoting, executing, and verifying machine-facing APIs, tools, agents, and specialist computations.

A catalog for discovering, comparing, quoting, executing, and verifying APIs, tools, agents, and specialist computations.

A zero-knowledge secrets manager that lets Claude Code, Claude Desktop, and Codex use API keys without exposing their plaintext to the agent.

Tanovar reviews payment-platform guides, API contracts, schemas, sample code, and test material to find contradictions that can derail developer integrations.

A behavioral-genetics researcher’s site pairing published work with PaperPull, an AI-assisted tool for turning research papers into structured meta-analysis data.

A local tool that turns supported OpenAI-compatible traffic into cost receipts attributed to customers, workflows, or tasks.

VeraTracer is a checkpoint-based data-workflow tool whose VeraScript language lets users branch runs and return to earlier analysis steps without recomputing everything.

A deployment platform that imports vibe-coded apps, scans and builds them, provisions Postgres, adds HTTPS, and publishes them at a live URL.

A native Windows and macOS desktop client for database development, administration, SQL editing, monitoring, and schema work.

A Portuguese-English product site for controlling and monitoring industrial power supplies, chargers, and UV LED systems through CAN telemetry, touch interfaces, and remote access.

A Japanese open-source project developing SwiftUI apps, on-device and generative AI, models, and datasets.

An independent software lab building tools for keeping personal logs and photos close at hand, then passing selected material to a destination the user chooses.

A shared graph of Markdown notes that AI agents can mount over MCP.