Havuz

Plan, apply and roll back changes to a production ontology, with a blast radius report and a proof an auditor can re-check.

Open decision models from 0.8B to 27B: a calibrated probability for every option in one pass, and a short thought only when unsure. Engine f

Neuro-symbolic agents with verified memory. MCP math tools, an algebra repair Skill, and an open challenge. Verify solutions, learn checked

Transforms Databricks Unity Catalog tables into a dynamic knowledge graph, with ontology design and reasoning exposed as tools via MCP.

[Public preview] Hard human-written reasoning puzzles: 749 answered items in 15 topics (707 scored core), with protocol and scorer.

Reasoning-effort editing for third-party models in DeepSeek Harness: per-model thinking levels with a knowledge base + protocol inference, e

风插 2.0 — DSH 推理语域锚定预设：两个 agent preset（风插标准/风插 PTC）在 runtime-context 快照头尾各注入一次形式化约束，使推理以第一人称复数展开。定性=硬指令约束（非吸引子）；未证明能力提升，未证明无负面作用。Apache-2.0

TTPO: Test-Time Policy Optimization

为 DeepSeek Harness 打造：选中 AI 回答或思考内容，随时追问、学习、检查与纠偏。

DeepSeek Harness (dsh) 插件：注入系统提示词，强制 Agent 的思维链 / 规划 / 工具推理全程使用简体中文，防止中文思考漂移为英文。零外部依赖，任意本地目录 link 即装。

Qwen3.8-Flash-Next 177B MoE (IQ3_S 3.44bpw) on RTX 5090 Laptop 24GB: full-model quality @ 97.7 tok/s · dual-lane 256K full-context unlocked

Codex ile yerel kod inceleme ve testleri korurken, karmaşık teknik akıl yürütme görevlerini ChatGPT Web'e aktaran deneysel bir araçtır.

Neden yükseliyor · 28 günlük yaşında 142 yıldız dikkat çektiğini gösteriyor; aynı dönemdeki 12 commit geliştirme faaliyetine işaret ediyor. Ancak commit'lerin tek yazardan gelmesi, bağımsız katkı veya kullanım kanıtı sunmuyor; yıldız verisi de tek başına benimsenmeyi doğrulamıyor.

MathCode: A Frontier Mathematical Coding Agent

A framework for building intelligent agents

Evidence-graded, agent-executable thinking-method skills - 63 frameworks reduced to their working mechanism, honestly graded, and producing

Kiro provider for DeepSeek Harness with Builder ID login, live model discovery, and reasoning effort controls

NeuroArxiv, Claude Code ve Codex CLI için bir skill/CLI olarak arXiv makalelerini tarayıp teknik tasarım önerisini kaynaklara dayandırmaya çalışır.

Neden yükseliyor · 53 günde 433 yıldız alması ilgi gördüğünü gösteriyor; arXiv araştırmasını agent iş akışına eklemesi ve README'deki karşılaştırmalı eval iddiası dikkat çekmiş olabilir. Ancak eval, README'de beş problemle sınırlı ve bağımsız doğrulama bilgisi sunulmuyor; son 28 günde 1 commit ve 1 farklı yazar görülmesi de bakım/katkı ivmesinin zayıf olduğuna işaret ediyor.

A reasoning skill for AI agents to improve complex judgment, premise auditing, counterargument pressure, and verifiable conclusions.

Open-source middleware for AI agents — catches mid-run failures,compresses stale context, and grows a reasoning library across runs. Any mod

Mental-R1: Aligning LLM Reasoning for Mental Health Assessment

Restores reasoning content dropped from multi-turn OpenCode conversations (DeepSeek, Kimi, GLM, MiMo, MiniMax, InclusionAI/Ling, OpenRouter)

Method for Long Context RLMs using verifiable Lambda Calculus

🌟 推理王国：关于 AI 推理机制的思想实验手册。从信息论、符号逻辑与表示学习出发，系统剖析大模型“智能”的本质。

Distill teacher chains-of-thought into a LoRA adapter via a strict boxed-answer format contract + two-phase Train→Nudge (silver-medal NVIDIA

Perplexity AI search, reasoning, research, and compute - MCP server, dashboard, and multi-IDE auto-config for VS Code.

A Complete faster runtime for coding agents. Optimize the journey, not the hop. Make coding agents 25% faster and 30% cheaper on average whi

CTX is the standard Cognitive Version Control System for AI. Structured working memory for AI agents. CTX preserves goals, tasks, evidence,

An Agent with Semantic Memory. This is an experiment for Neuro-Symbolic AI

Persistent organisational intelligence for AI agents. Turn codebases, documents, policies and decisions into a knowledge graph — 13 LLM back

[IJCAI2026] Official code for paper "Cross Domain Test Time Scaling: Scale Knowledge and Reasoning on Cross Domains"

A local-first cognitive architecture for building a persistent digital companion with memory, reasoning, personality, and long-term learning

Doubt, until what remains is certain. An MCP server that recursively doubts every plan decision — grounding each in real evidence or asking

Reformat a live LLM token stream inflight — separate thinking from answer, collapse repetition, normalize. Google Vertex / OpenRouter / meta

Think only when the tool call looks unsure: confidence-gated reasoning for small local models. Code, BFCL results and paper.

Experimental DSH Minimal trajectory-steering adapters for DeepSeek V4 Pro on Pi, Oh My Pi, and OpenCode V2. 推理轨迹引导实验。

Use GLM models in GitHub Copilot Chat with your own API key. Supports image input, tool calling, and streamed thinking.

Qt/C++ OpenAI-compatible reasoning degradation guard proxy with GUI, CLI, deb and AppImage packaging

AI reasoning skills distilled from 4,665 real Claude Fable 5 chain-of-thought traces. Mathematically tuned. Grade A (100%) emulation accurac

Real-time chain-of-thought anchoring monitor & intervention plugin for DeepSeek Harness: three-band (spec/mixed/react) fingerprint detection

SciPIvot: Benchmarking Principal-Investigator Judgment at Pivotal Moments of Scientific AutoResearch Agents

Ground LLM output as provenance-tracked claims, not truths. A maximal super-ontology + native RDF-1.2 logic: reasoning engine — OWL, SHACL,

Open-source thinking-framework skills that make rigorous reasoning executable for AI agents — first-principles, inversion, second-order thin

Raw benchmark logs for Swift-Qwen3.8-27B vs the Qwen3.8-27B BF16 base: every response, score, and config across 9 benchmarks