-
SkillOpt proposes an executive strategy for agents that decide which skills to evolve and keep
SkillOpt, from Microsoft Research, is the top-upvoted agent paper on Hugging Face today. It tackles a layer above individual skills: the executive strategy for a self-evolving agent — deciding which skills to create, refine, keep, or discard as the agent accumulates experience. ## The self-evolving-skills space A growing body of work frames the same closed… Continue reading
-
stop-slop is a skill file that strips the AI tells out of your prose
stop-slop is a Claude Code skill with one job: catch and remove the patterns that make AI writing recognizable as AI writing. Add it as a skill, drop it in a Claude Project, or paste it into an API system prompt, and it rewrites against a specific list of tells. ## What it bans Throat-clearing… Continue reading
-
ECC is an agent-harness performance optimizer with a built-in security scanner for MCP servers and tool permissions
ECC is an agent-harness performance optimization system — skills, instincts, memory, security, and research-first development — that works across Claude Code, Codex, Cursor, OpenCode, Gemini, Zed, and GitHub Copilot. It exploded on GitHub trending with 2,000+ stars in a day. Built originally at the Claude Code Hackathon (Cerebral Valley x Anthropic). ## AgentShield The standout… Continue reading
-
Yansu watches how you work and builds custom apps before you ask — proactive AI from Isoform
Yansu, from Isoform, is a proactive AI app builder. Instead of waiting for a prompt, it observes how you work — via desktop screenshots and messaging apps — learns your patterns, and builds custom applications, workflow improvements, bug fixes, or insight dashboards on its own. ## The proactive angle This is the same “proactivity” frontier… Continue reading
-
MashuPack compiles the exact parts of a repo into one clean file for ChatGPT and Claude — no upload, no backend
MashuPack solves a small but constant annoyance: getting the right slice of a codebase into a browser AI chat without messy copy-paste or hitting file limits. Select exact files or subsystems from a local repo, and MashuPack compiles them into one structured text file ready to paste into ChatGPT or Claude. ## How it works… Continue reading
-
WorldKV lets world models remember what they have seen — training-free KV retrieval and compression for spatial consistency
WorldKV, from KAIST AI and Naver AI Lab, tackles a core problem in world models: when you revisit a place you’ve already seen, the model should show you the same thing. Sustaining that persistent consistency has been hard — full attention preserves it but blows the real-time budget; sliding-window inference is fast but forgets. ##… Continue reading
-
Anthropic-Cybersecurity-Skills maps 754 security agent skills to MITRE ATT&CK, NIST, and 3 other frameworks
Anthropic-Cybersecurity-Skills is an open-source library of 754 structured cybersecurity skills that give an AI agent the workflows of a senior security analyst. It hit GitHub trending with 930 stars in a day. Apache 2.0, spanning 26 security domains. ## The framework mapping What sets it apart: every skill is mapped to five industry frameworks —… Continue reading
-
PhysX-Omni generates simulation-ready 3D objects — rigid, deformable, and articulated — for embodied AI
PhysX-Omni is a unified framework for generating simulation-ready physical 3D assets across object types: rigid bodies, deformable objects, and articulated objects (things with joints, like doors or robot arms). “Simulation-ready” is the key phrase — the outputs aren’t just pretty meshes, they carry the physical properties a simulator needs. ## What’s under the hood A… Continue reading
-
Salesforce Agentforce Coworker puts a CRM agent in the search bar — Slack, Teams, and ChatGPT too
Salesforce announced Agentforce Coworker — an AI agent that lives in the search bar and, in its words, “doesn’t just find things, it works with you.” Ask it anything in plain language from Salesforce Global Search and it taps CRM data, workflows, customer history, opportunities, and cases, then takes action in real time. ## Where… Continue reading
-
earendil-works/pi is an AI agent toolkit — coding CLI, unified LLM API, TUI libraries, and vLLM pods in one repo
pi is an open-source AI agent toolkit that bundles a lot into one repo: a coding agent CLI, a unified multi-provider LLM API (OpenAI, Anthropic, Google), TUI and web UI libraries, a Slack bot, and vLLM pods for self-hosting. ## The unified-API core At the center is a single API that abstracts across providers —… Continue reading
Top AI Product
Every day, hundreds of new AI tools launch across Product Hunt, Hacker News, and GitHub. We dig through the noise so you don't have to — surfacing only the ones worth your attention with honest, no-fluff reviews. Explore our latest picks, deep dives, and curated collections to find your next favorite AI tool.
Top AI Product 