Developer Corner

Fictionet describes a controlled network for testing how AI agents behave when they encounter misleading sites, poisoned data, or unreliable systems.

A set of focused Rust crates for building 2D and 3D jam games that can run natively or in a browser from the same code.

An open-source agent skill uses ADB to disable preinstalled Android apps without root or uninstalling them, with backups and an undo path.

CmdID analyzes command lines and presents documentation-based explanations alongside abuse references and security detections.

A no-login sandbox runs AI agents against sample data and shows which actions its policies allow or block.

A local CLI that lets developers retrieve recorded AI-agent prompts and reasoning for a line of code.

Loupe is a Video.js v10 skin that turns a video’s storyboard frames into a clickable contact sheet, with an option to save the sheet as a linked PDF.

Silkscreen is a free, offline macOS bench tool that combines hardware design files and documents into an assembly record for tracing connections, reviewing conflicts, and running bench checks.

A benchmark report tests eight models on 15 game-development tasks across nine Unreal Engine games, scoring whether the running games pass deterministic checks.

A pre-alpha layer that sits between AI agents and business systems, journaling writes and offering conflict-checked undo.

Free, MIT-licensed firmware for the Baofeng DM-32UV adds an FM transceiver and a collection of radio apps, with flashing available through a browser.

An open-source tool lets Claude Code, Codex, and other MCP agents message one another across machines and accounts.

An open-source web app for Japanese startups to manage shareholder records, cap tables, stock options, and company meetings from one ledger.

offweb fetches text from Gemini, Gopher, and IPFS URLs for agents, then returns a signed receipt describing the retrieval.

A bilingual portfolio for Régis Rigaud’s independent AI workshop, listing eight public tools and sixteen projects in progress.

A desktop app for running coding agents in isolated sandboxes on your computer or your own server, then organizing the work into repeatable workflows.

A desktop VFX generator creates spell and impact effects for games, with exports for Godot, Unity, and sprite sheets.

A Spanish-language reference for PLC, drive, and HMI faults, with troubleshooting guides, calculators, a glossary, and a technical forum.

A searchable catalog of power-electronics and control papers, tracking reproduction results, evaluation status, and gaps in the source material.

An open-source CLI lets shell-capable coding agents email a person for a decision, receive replies, and handle signup verification codes.

EDIParse presents a browser-based workbench for inspecting X12 EDI files and converting them into accounting exports, retailer documents, and shipping labels.

A game studio offering games and mods designed for blind and sighted players to play together or solo.

A browser-based simulator for exploring CAN and CAN-FD bus topologies, signal reflections, receiver bit sampling, and power.

A beta service for agents that recomputes engineering formulas and arithmetic claims, then returns a signed result link.

CycTronics is developing a USB hub and attachable modules for load cells, relays, and audio, with a shared software interface.

A pre-alpha control layer that previews Kubernetes changes from AI agents, applies policy, requests approval for risky actions, and grants a credential limited to the approved change.

A free, no-signup sandbox for testing SMART on FHIR apps against made-up patient records.

A machine-oriented service presenting listed-company research pages built from issuer records, with source references, read dates, and a receipt chain.

biorun offers browser, API, and coding-agent access to published biology foundation models, with a baseline and execution details for each run.

An engineering project publishing native 64-bit ports of classic PC strategy and role-playing games, with Linux, ARM64, handheld, and cooperative-play support.

A platform for giving AI agents cloud phones to inspect and control mobile apps.

A European imbalance-data API offers prices, volumes and nowcasts, alongside source attribution, latency measurements and revision history.

A gallery of kiosk widgets for monitoring AI coding sessions, machine vitals, calendars, weather, usage, and news.

An independent security research site covering vulnerabilities in Android, mobile, IoT, and embedded systems.

Opsis records actions on a real website or terminal, then turns the run into product-demo videos.

Phone Use is a desktop tool that lets AI agents view and interact with an iPhone, including inside apps without APIs.

A searchable reference catalog listing network devices with port details, dimensions, power information, YAML files, and front- and rear-view drawings.

It presents a payment-control layer that checks an agent’s transactions against a signed mandate defining limits, payees, and expiry.

An open-source local CLI reads Claude Code logs, estimates where token usage went, and ranks possible fixes.

A free, open-source Windows utility for finding and cleaning up stale developer processes, AI-agent sessions, caches, and resource-heavy WSL/Docker workloads.

A tool that lets Claude read, annotate, and navigate reMarkable Paper Pro and reMarkable 2 tablets.

SecureCrawl says it connects to WordPress hosting over SSH, compares core and plugin files with official releases, and checks the database for unexpected content.

A hardware controller and cloud portal for restarting, monitoring, and securely reaching equipment at remote sites.

A home-network scheduler that sends GPU jobs from cloud apps to compatible machines with available memory.

The site presents a research project for observing how AI crawlers and browser agents find web pages, handle controlled tasks, and identify themselves.

TXGR lets authors write an agent skill’s workflow in plain text, then maps and checks its declared steps.

A small arcade designed to run inside OpenCode Desktop and Antigravity while an AI agent works.

A beta project for recording who helped whom, how much was contributed, and whether it was later repaid, with a hash-linked ledger and public verification materials.

Post Markdown presents a plain-text memory service where AI agents save Markdown notes through an API and people can read and restore their versions.

QRSig is a draft format for embedding a signature in a QR code so compatible readers can check who signed its contents.

An open-source project for account-free, Tor-routed chat that says messages stay end-to-end encrypted and are not stored on a server.

A browser-based workflow for planning a game, trying a model-built version, logging a playtest, and working toward the next version.

Treehead presents a searchable feed of TLS certificates and a WebSocket stream for newly issued certificates.

An answering-machine-detection API that classifies the first seconds of outbound calls as human, voicemail, IVR, or another outcome.

Bitsettled is a board where merchant offers and buyer wants become threads for personal agents to negotiate.

An interactive Stripe Issuing demo walks through an AI agent receiving a limited allowance, making permitted purchases, and encountering a declined one.

AgentsChecked presents BlameGraph, a tool for tracing field changes across multi-agent AI runs and identifying the step where a value changed.

A mountain-themed MCP and REST service lets Claude-connected clients or standalone bots join tables, play full games, and answer Bamboo-Stick challenges.

PayCall describes a system where agents pay for HTTPS APIs using signed vouchers, with seller earnings paid out through PayPal.

AcctRelay is building a Windows agent and cloud API to connect SaaS products to customers’ Sage 50 desktop accounting data.

A browser-based node editor for configuring and deploying Raydium LaunchLab platforms, with settings for curves, fees, liquidity, vesting, and Token-2022.

An API for recording AI-agent runs, trace spans, and scores, with redaction defaults and usage-based event quotas.

A pilot MCP server designed to check Texas statutory lien waivers against payment, job, vendor, and commitment records.

LLMFreeAPI presents a free, keyless OpenAI-compatible API for testing model refusals, alongside a buyer’s checklist.

A software platform for discovering sensor feeds, combining partial observations, and producing detections and forecasts at the edge.

The page describes a SvelteKit deployment platform that bundles HTTPS, a per-app Postgres database, mail, and a local LLM.