What should a search agent learn from when deciding whether to search? 🔎 Search When Needed, Proceed When Reliable: Metacognitive Reliability Supervision for Efficient Agentic Search ❓ What goes wrong when the reward is just final-answer correctness? 💡 A lucky guess, a correct answer with needless searches, and a wrong answer after useful retrieval all look alike, so the reward says little about each search decision. That leads to under-searching on unreliable states and over-searching on reliable ones. ❓ What does this paper reward instead? 💡 The estimated reliability of the intermediate conclusions the model writes at each step. Uncertainty is the mean negative log-likelihood of the tokens, and a mapping fitted beforehand on sampled data turns it into a reliability score. ❓ How are search and no-search decisions scored? 💡 ・If the agent proceeds without searching, the score is its current reliability minus a calibration reference ・If it searches, the score is reliability with the retrieved evidence minus reliability without it The average becomes a process reward added to GRPO. Probing is used only in training, so inference has no extra cost. ❓ How well does it work? 💡 With Qwen2.5-7B it averages 69.3 CEM over seven QA benchmarks, versus 67.2 for the strongest baseline, HiPRAG, with 1.64 searches per question (6.8% fewer). Multi-hop searches rise about 6.6%, and training is about 14.6% slower per step. #AgenticRAG# #ReinforcementLearning#

The open-source WeKnora (MIT, 22k stars) just shipped v0.8.0. A new step from RAG toward agentic RAG: a self-remembering, safely sandboxed knowledge agent platform. What's new: 🧠 Cross-session memory: profile, preference, fact, task and interest, each proposed for your confirmation before it sticks. 🔐 Sandboxed by default: Skills used to run in the host process; that backend is gone. Each session now runs in its own sandbox on Docker, E2B, or Cube, with configurable network egress policies. 📦 Skills as assets: a tenant-wide catalog, not scattered scripts. Pull them from ClawHub, SkillHub or any git repo, versioned and verified to load before anything runs.

TRM is proud to be a to partner on the Cambridge Digital Innovation and Regulation Initiative’s ✨Global Agentic Regulator Hackathon✨ (July 8 – September 18, 2026). AI is outpacing public sector adoption, and the risks are growing. This hackathon brings global experts together to build trusted, deployable solutions. Open to: ▪️ Regulators ▪️ AI researchers ▪️ Academia ▪️ Financial institutions we ▪️ RegTech & SupTech Learn more and submit a concept note by August 9 at 12:00pm BST: @CambridgeAltFin #AgenticAI# #RegTech# #SupTech# #DigitalInnovation# #AIforGood# #CDIRHackathon# #AgenticRegulator# #AgenticFinance#

In an era of #agenticAI#, documentation 📚must also adapt🤖 I went from spending 100k tokens to just 2k tokens using graphs, a 95% savings🤩 This is a project I did on @chainlink using CRE,CCIP,ACE where the main code section is a for directory of 4,650 files (7 million words)💻

We're heading to San Jose for AGNTCon + MCPCon North America (@AgenticAIFdn), October 22-23. If you're building agents on MCP or getting them into production, come find us. Event details:

Wow... this is genius. AgentCraft is a team of Claude agents doing real work on your code, inside a Minecraft studio you can walk around in. Super cool. →