Last 90 days
Topical activity
These are the most active topics in the published evidence. An item can match more than one topic.
Models and Ecosystem · 1120 evidence items
Open-source software is distributed with source code and an open-source licence. Check the model weights, model licence and hosted service separately.
Recent coverage includes agent-device GitHub Overview: Semantic Mobile Verification for AI Coding Agents and Work together, your way, with GitHub Copilot.
The last 90 days contain 1120 matching stored items from public github discoveries, public web article, public youtube video.
Latest matching evidence: 2026-11-07 T16:00
Production Engineering · 858 evidence items
Agent evaluations measure whether a system completes representative tasks safely, correctly and consistently.
Recent coverage includes I Turned a 6GB Laptop GPU Into a Local AI Coding Agent and I Turned an Old PC into a Worker for My AI Agents.
The last 90 days contain 858 matching stored items from public github discoveries, public web article, public youtube video.
Latest matching evidence: 2026-11-07 T13:51
Capabilities and Interfaces · 730 evidence items
Tools let agents take bounded actions or retrieve information; MCP standardises how many tools and resources are exposed to clients.
Recent coverage includes OutSystems MCP Explained: Connect AI Coding Agents to Your Entire App and Part-3: Connect REST API to MCP Server with Node.js | GitHub Copilot in VS Code 🚀.
The last 90 days contain 730 matching stored items from public github discoveries, public web article, public youtube video.
Latest matching evidence: 2026-11-07 T13:30
Production Engineering · 560 evidence items
An agent harness is the application layer that manages models, tools, state, limits and execution behaviour.
Recent coverage includes agent-device GitHub Overview: Semantic Mobile Verification for AI Coding Agents and The Rise of the Super-Agent: Deep Seek + Claude Code.
The last 90 days contain 560 matching stored items from public github discoveries, public web article, public youtube video.
Latest matching evidence: 2026-11-07 T16:00
Capabilities and Interfaces · 496 evidence items
Skills package repeatable instructions, resources and workflows so an agent can apply specialised capability consistently.
Recent coverage includes Anthropic Just Dropped Claude Personal Finance (Full Breakdown) and 7 NEW Claude Mods Use Cases (Try These Now).
The last 90 days contain 496 matching stored items from public github discoveries, public web article, public youtube video.
Latest matching evidence: 2026-11-07 T15:07
Governance · 484 evidence items
Safety and security engineering limits harmful, unauthorised or compromised behaviour across models, tools, data and agent runtimes.
Recent coverage includes Lumi: A Desktop Agent That Uses the AI You Already Have and DEF CON 34 - AI - AgentBreaker A blind spot detector for coding agents - A Narasimhan, F Kaiyom.
The last 90 days contain 484 matching stored items from public github discoveries, public web article, public youtube video.
Latest matching evidence: 2026-11-07 T00:00