A developer built Ofrenda, an offline command-line tool that interviews users about a person they want to remember and writes two plain Markdown files: a short remembrance and a checklist of items for…
A developer published a companion example and course material arguing that agent safety must be enforced in the application harness rather than the model, using the principle "The model proposes. The …
The SynuraCore team documented how its AI coding agent's save command returned exit code 0 while silently saving nothing, because the shell function it called was defined in ~/.bashrc behind an intera…
Maya Research released Maya 2 Native, an Apache-2.0 text-to-speech model covering 11 Indic languages plus English, with 20 built-in voices, streaming audio, and voice cloning from 3–10 seconds of refe…
A developer rebuilt a desktop AI coding agent's shell-command auto-approval system from a denylist to an allowlist after finding four exploitable loopholes: command chaining, argument injection, envir…
A developer published a 35-question AI vendor and agent security questionnaire that extends standard vendor reviews beyond data-hosting and SOC 2 checks to cover what an agent can actually do, includi…
A developer at Axonscape's Launchpad documented a common deploy failure in which builds break with "Could not resolve" errors for imports whose capitalization doesn't match the actual filename, even t…
A developer built a test harness, using Claude Code, that evaluates whether an AI sales agent correctly routes inbound emails to either an agent-drafted reply or a human, based on a six-line JSON rule…
A developer with two decades of programming experience used a week of AI-assisted coding to build CraftEduCenter, a browser-based learning playground spanning twenty subject areas, including molecule …
A developer demonstrated a durable-execution wrapper for CrewAI agents that converts each tool call into a Dapr workflow activity, so a crashed or redeployed process resumes from the last completed ca…
A developer maintaining the twd-ai Claude Code plugin described building an eval suite with Claude Code's `claude plugin eval` command to test agent skills rather than relying on manual spot checks. E…
A developer built Bookshelf Séance, an open-source tool that ingests a user's EPUB library and lets them ask questions of a chosen author, with the ghost answering in the first person using only passa…
A developer published a hands-on guide to using Ragas 0.4.x, an Apache-2.0 open-source library for evaluating RAG and LLM applications, to replace subjective "vibe checks" with trackable metrics. The …
A blog post outlines a four-phase migration to post-quantum cryptography, warning that nation-states are already intercepting and storing encrypted traffic for later decryption under a "Harvest Now, D…
A developer has released FeedLog, a self-hostable AI customer-support and feedback-board product that runs on Docker, Vercel, or Cloudflare Workers. The project uses OpenAI's text-embedding-3-large at…
An engineer outlined a design approach for audit trails in agentic healthcare AI, arguing that logging only final answers is insufficient when agents retrieve records, call tools, update systems and d…
A developer tested Spark-X2.6-4B, a roughly 4.11B-parameter Apache 2.0 model from SparkLLM/XHToken, on a 16GB MacBook Pro using MLX Serve, finding it among the top two or three small models tried on t…
A developer outlined a review protocol for coding agents that treats a green test check as a coverage claim rather than proof of correctness, since results can come from a stale commit or a suite that…
A developer behind the open source browser-agent project described how multiple AI agents sharing a single logged-in Brave browser repeatedly wiped out human work by reloading pages, duplicating tabs …
A developer built EchoArchive, a local-first application that preserves spoken oral histories by treating the original recording and human-reviewed transcript as the sole authorities, with local AI us…