Research

17 papers, memos and threads on agent architecture. These are reading notes, kept separate from the directory on purpose: they are not frameworks, they carry no score, and they are never counted among evaluated tools.

• analysis · added 6 Apr 2026

  Code Review Agents Empirical Study (arXiv:2604.03196): Deep Technical Analysis

  This is an empirical study, not a system. It analyzes pull requests from the AIDev dataset (HuggingFace), classifying reviewer compositions and evaluating signal quality of automated code review agent (CRA) outputs.

• paper · added 6 Apr 2026

  Paper Analysis: Agentic Federated Learning

  A position/vision paper (authors' own framing: *"we demonstrate the viability of integrating LM-Agents into FL with a proof-of-concept"*) proposing Agentic-FL: replacing static federated learning coordination protocols with LLM-based agents. The central argument is that existing FL solutions address isolated concerns (client selection, aggregation, privacy, communication) with fixed algorithms that cannot adapt to the dynamics of real FL environments. The paper claims LM-agents enable holistic, simultaneous management of all these concerns through contextual reasoning.

• analysis · added 4 Apr 2026

  ARIS: Auto-Claude Code Research in Sleep — Deep Analysis

  ARIS is a methodology-first, Markdown-driven skill system for autonomous ML research workflows. It orchestrates cross-model collaboration — Claude Code executes research while an external LLM (Codex, Gemini, or other) reviews work as an adversarial critic. The entire system is files + plain Markdown skills (no database, no framework), making it portable across Claude Code, Cursor, Trae, Codex CLI, and other agents.

• paper · added 4 Apr 2026

  arXiv:2604.02226 — When to ASK: Uncertainty-Gated Language Assistance for RL

  The paper addresses a real design question — when to seek help — but from the wrong direction for Forge. Our stack is LLM-native; MC Dropout cannot be applied to transformer inference.

• memo · added 3 Apr 2026

  ARIS (Auto Research In Sleep)

  ARIS is a prompt-engineering framework where the LLM *is* the runtime. 5 MCP servers bridge external LLMs for cross-model review, 5 CLI tools handle arxiv/Semantic Scholar fetching + GPU watchdog monitoring, and 49+ Markdown "skill" modules define composable research workflows (with YAML frontmatter) consumed directly by Claude Code/Codex/Cursor. The core architectural insight — "the LLM doesn't need a scheduler; the LLM IS the scheduler" — is orthogonal to Forge's programmatic orchestration but surfaces several quality patterns worth stealing: cross-provider adversarial review, provider-specific parameter clamping, thread history persistence for stateless APIs, and private dotfile API key fallback.

• paper · added 3 Apr 2026

  arXiv:2604.02155 — Brief Is Better: Non-Monotonic CoT Budget Effects in Function-Calling Agents

  The paper delivers an unexpected but well-supported finding: function-calling agents should think briefly, not deeply. The optimal CoT budget for tool selection is 8–16 tokens — approximately one sentence identifying the function and key arguments. Beyond that, reasoning quality degrades through a documented "dual failure" mechanism where extended thinking causes both function hallucination (the model generates names outside the candidate set) and wrong-function selection (the model talks itself out of the correct choice).

• post · added 14 Mar 2026

  @varun_mathur — Autoquant v2.6.9 Tweet

  > Autoquant: a distributed quant research lab | v2.6.9

• memo · added 9 Mar 2026

  Agentic Critical Training (ACT)

  ACT is a two-stage RL training paradigm that fixes the fundamental weakness of imitation learning (IL): IL teaches agents *what* expert actions look like, but never forces models to understand *why* those actions are better than alternatives.

• paper · added 9 Mar 2026

  arXiv:2603.03329 — AutoHarness: Improving LLM Agents by Automatically Synthesizing a Code Harness

  AutoHarness tackles a critical LLM agent failure mode: agents making illegal/invalid actions.

• paper · added 9 Mar 2026

  arXiv:2603.10062 — Multi-Agent Memory from a Computer Architecture Perspective

  The paper's single core claim: multi-agent system reliability is a memory problem, not a compute problem. The bottleneck in collaborative LLM agent systems looks "surprisingly familiar to computer architects" — it is the same memory hierarchy, bandwidth, and consistency challenge solved in CPUs, transplanted to semantic agent context.

• paper · added 9 Mar 2026

  arXiv:2603.10165 — OpenClaw-RL: Train Any Agent Simply by Talking

  OpenClaw-RL is a live, online RL training framework that trains language model agents *during production use* by extracting learning signals from the natural next-state feedback that already exists in every agentic interaction: user replies, tool outputs, error traces, test results, environment state changes.

• paper · added 9 Mar 2026

  arXiv:2604.02318 — Stop Wandering: Efficient VLN via Metacognitive Reasoning (MetaNav)

  ### Rubric Scores

• memo · added 9 Mar 2026

  From Hallucination to Structure Snowballing: The Alignment Tax of Constrained Decoding in LLM Reflection

  This paper is a diagnostic study, not a system paper. It investigates a three-component agentic self-correction pipeline built on the Reflexion framework:

• thread · added 9 Mar 2026

  HN Multi-Agent Framework Link Triage

  47 unique URLs extracted across 6 categories from 6 HN threads (1,100+ combined points, 418 comments). The HN multi-agent community is skeptical of framework proliferation but hungry for:

• analysis · added 9 Mar 2026

  Market Research: AI Agent Orchestration Platforms

  The AI agent orchestration market has exploded from $5.25B (2024) to $7.84B (2025), projected to reach $52.62B by 2030 (46% CAGR). The landscape is consolidating around 4 tiers: hyperscaler frameworks (Google ADK, Microsoft Agent Framework, OpenAI Agents SDK, AWS Strands/AgentCore), open-source orchestrators (LangGraph, CrewAI, Agno, PydanticAI, Mastra), protocol standards (MCP, A2A, Agent Skills), and specialized/research frameworks. >40% of agentic AI projects risk cancellation by 2027 due to cost/complexity — the gap between experimentation and production is the central market opportunity.

• memo · added 9 Mar 2026

  Trajectory-Informed Memory Generation for Self-Improving Agent Systems

  Agents repeat mistakes. They fail the same way, miss optimization opportunities, and don't transfer successful strategies across tasks. This paper proposes a 4-component pipeline that automatically extracts structured, typed learnings from execution trajectories and injects them into future agent prompts via similarity-based retrieval.

• analysis · added 9 Mar 2026

  Trajectory-Informed Memory Generation for Self-Improving Agent Systems — Technical Analysis

  LLM agents are amnesiac: they repeat the same failures, miss reusable successful strategies, and cannot automatically apply lessons from past executions. Existing approaches are inadequate: