Catalog

This is the main catalog of the wiki.

Overview

• home - Top-level overview of the knowledge base and workflow.
• home-zh - Chinese homepage for the public wiki dashboard.
• home-ja - Japanese homepage for the public wiki dashboard.
• overview - Structural map of the repository layers, sections, and active public domains.
• log - Chronological record of ingests, analyses, queries, and maintenance actions.
• Website publishing layer - site/ builds the public wiki/ layer with Quartz and GitHub Pages while excluding private/raw materials.

Entities

• canvas-lms - Canvas coursework LMS note; current user-provided outage/hack status blocks syllabus retrieval for UMN transfer-credit review.

• yao-shunyu-alfred - alfredyao public corpus for physics-to-AI, Google DeepMind/Anthropic, quantum physics, RL, and agentic coding.

• yao-shunyu-ysymyth - ysymyth public corpus for language agents, OpenAI work, ReAct, ToT, SWE-agent, SWE-bench, tau-bench, CUA, and Deep Research.

• lidang - Public idea-stream corpus for Lidang / lidangzzz, focused on YouTube, X, path engineering, AI coding, and practical heuristics.

• andrej-karpathy - Public AI researcher/educator corpus focused on GitHub projects, writing, videos, and reusable technical heuristics.

• vipin - Central person page for the owner/subject of this knowledge base.

• university-of-minnesota - Issuer entity for University of Minnesota housing contract materials.

• analog-agent - AI4EDA layered analog circuit design workbench entry with SPICE/configured-truth boundaries, startup packet, module map, and artifact safety rules.

• uncertainty - Current main project for task-grounded calibrated uncertainty in LLM-based recommendation.

• truce-rec - Uncertainty-aware generative recommendation project with strict evidence labels and official-baseline gates.

• tgl-rec - Temporal graph-to-language recommendation project focused on next-need transitions.

• donebench - Specification-grounding benchmark for tool-using agents.

• uncertaintyprotein-ai4s - Exact local project-name entry for the AI4S protein optimization feedback-shift project.

• protein-optimization-feedback-shift - Conceptual project page for uncertainty-aware closed-loop protein sequence optimization.

• weipingyan-portfolio - Local portfolio workspace containing public site/profile repos and a private memory site.

• academic-portfolio - Sensitive academic/application archive.

• undergraduate-projects-netherlands - Undergraduate project archive for Netherlands/TU/e coursework projects.

• undergraduate-study-netherlands - Undergraduate study archive organized by course/domain.

• qmq-app - Saved tool reference for US visa appointment slot monitoring / “qiang mei qian” workflows.

• medora - Local healthcare project for AI-native personal health records and healthcare workflows.

• anbeime-skill - Local skill-store and skill-library repository under D:/Skill.

• colleague-skill - Local dot-skill repository for distilling people/context into reusable skills.

• darwin-skill - Local skill optimizer inspired by autoresearch-style evaluation loops.

• mattpocock-skills - Local mirror/fork of composable engineering and productivity agent skills.

• nuwa-skill - Local repository for research-driven person-perspective skill creation.

• paper-orchestra - Project-local PaperOrchestra skill pack for multi-agent research-paper writing and installed Codex skills.

• content-creation-publisher - Project-local Anbeime content creation/publishing skill pack for web capture, Markdown formatting, illustration, and WeChat/X posting.

• wechat-video-channel-publish-skill - Local Playwright CLI skill for WeChat Channels / 视频号 login, cookie check, video upload, covers, captions, scheduled publishing, and draft upload.

• frontend-design - Project-local Anbeime frontend design skill for distinctive, production-grade interface aesthetics.

• chrome-automation - Project-local Anbeime Chrome automation skill using agent-browser and real Chrome CDP control.

• lark-cli - Official Lark/Feishu CLI runtime and installed lark-* skill set for Docs, Wiki, Drive, Base, Sheets, IM, and related API workflows.

• feishu-bridge - Project-local Feishu/Lark router skill that uses lark-cli for API access and chrome-automation for web-only links/forms.

• readme-blueprint-generator - Project-local README specialist skill for high-quality README generation, critique, and rewrite workflows.

• agent-resources - Shared D-drive skill/resource library and implicit skill-routing map for local agents.

• email-assistant-skill - Project-local Gmail/Google Workspace assistant skill for quiet mail triage, Chinese summaries, draft-first replies, and explicit send approval across UMN and vipinapple accounts.

• communication-assistant-skill - Project-local unified communication router skill for WhatsApp, WeChat, QQ, email, and shared outbox workflows.

• terraria-save-archive - Local Terraria save backup and inventory rooted at D:/Terraria_doc.

• lu-kan - Architect, exhibition designer, curator, and AIGC workflow practitioner profile from user-provided chat material.

Concepts

• llm-wiki - Persistent wiki pattern in which an LLM compiles knowledge into maintained markdown pages.
• llm-based-recommendation - Research area focused on using large language models within recommender systems and recommendation workflows.
• html-ppt-agent-workflow - Local D-drive workflow for generating HTML PPT decks with frontend-slides and beautiful-html-templates.
• feishu-material-access-workflow - API-first, browser-fallback workflow for connecting Feishu/Lark materials and shared forms to Codex.
• readme-maintenance-workflow - Rule for periodically refreshing the root README as wiki structure and automation workflows evolve.
• durable-agent-rule-memory - Rule that user-requested future-agent memories must be persisted into AGENTS/wiki rather than left only in chat.
• codex-automation-model-policy - Hard rule that cron automations use gpt-5.5 with xhigh reasoning when supported, not low or medium reasoning.
• agent-collaboration-tone-and-model-roles - User-facing collaboration rule for warmer partner language, Codex parallel selves, Opus/Sonnet roles, and DeepSeek Pro/Flash preferences.
• agentmemory-first-agent-collaboration - Active collaboration model using agentmemory for memory, signals, actions, and handoffs while keeping the wiki as public crystallization.
• implicit-skill-routing - Rule that agents should choose relevant skills from task intent and skill metadata, not only explicit user skill names.
• weiping-wiki-maintenance-system - Continuous WEIPING_WIKI maintenance system for whole-computer project routing, workstation-maintenance handoff, D-root move-with-junction organization, UUPF-aided skill upgrades, scoped auto commit/push gates, and tight-but-low-coupling agentic project links.
• weiping-agentic-project-constellation - Route map for WEIPING_WIKI, devtools, AGENT_RESOURCE, AGENTIC_SCIENCE/UUPF, WEIPING_LAB, and WEIPING_COUNCIL with alias, junction, artifact-handoff, and low-coupling rules.
• vipinknowledge-maintenance-system - Historical alias page that redirects old vipinknowledge / vipin-wiki maintenance links to weiping-wiki-maintenance-system.
• obsidian-feature-parity - Obsidian-compatible vault layer with Bases, Canvas, dashboard, commands, templates, backlinks, search, quick switcher, outline, preview, tags, properties, tasks, slides, and word-count commands.
• whole-computer-project-map - First-stop routing map for whole-computer maintenance across C/D/G drives, importance-based depth, workstation move boundaries, D-root organized buckets, active roots, archives, and safety boundaries.
• d-drive-project-map - First-stop routing map for D-drive agent infrastructure, public exports, private runtime state, AGENT_RESOURCE/AGENTIC_SCIENCE/WEIPING_LAB/WEIPING_COUNCIL routes, D-root organized buckets, and research isolation.
• local-active-project-roots - Detailed launch map for active non-research local project roots, including entry docs and safety boundaries.
• model-collaboration-context-and-reference-intake - Rule that partner handoffs and external reference searches must carry enough context and depth to be genuinely useful, not toy summaries.
• local-cc-sidecar-agent-workflow - Strict three-role Codex + Claude Code partner workflow with PixelCat/CC health check, upstream-credential failure handling, and read-only Opus/Sonnet prompts.
• agent-hub-mcp-server - Deprecated historical custom Agent Hub MCP server, superseded by agentmemory-first collaboration.
• agent-skill-installation-workflow - Durable rule for installing skills as usable tools: D-drive mirror, .codex/skills install, dependency setup, smoke tests, wiki docs, and scoped commit.
• research-ideation-policy - Rule for ambitious research ideation: no shallow stitching, force novelty, and allow radical reframing.
• research-project-workbench - Upper-level workbench route for analog-agent, Uncertainty/Pony-Rec, TRUCE-Rec, and TGL-Rec, including startup packets, claim gates, artifact boundaries, and collaboration priority.
• song-page-capture-rule - Rule that every liked song should get a dedicated clickable, playable song page when possible.

Sources

• 2026-05-31-voice-sample - Public voice sample archive with Hangzhou dialect, Mandarin, and English MP3 files.

• 2026-06-01-umn-az-hold-esl-course-information - User-provided verbatim University of Minnesota AZ hold, ESL course, English proficiency score, MN Battery Test, and MELP contact guidance.

• obsidian-dashboard - Generated Obsidian-style dashboard for bookmarks, top tags, frontmatter properties, and open checkbox tasks.

• obsidian-compatible-commands - Generated command-palette view for Obsidian-style local-first wiki operations.

• README - Generated home for note-derived Markdown slide decks.

• README - Generated Obsidian Bases index for all pages, active projects, source notes, and maintenance views.

• 2026-05-17-html-ppt-skill-and-deck-session - Chat/session source note for installing the D-drive HTML PPT skill/template setup and producing two validated decks.

• 2026-05-17-paper-orchestra-github - GitHub source note for Ar9av/PaperOrchestra, its paper-writing pipeline, local mirror, and installed skills.

• 2026-05-17-content-creation-publisher-skill - Source note for anbeime/skill’s content creation publisher skill, installed sub-skills, runtime notes, and usage patterns.

• 2026-05-17-wechat-video-channel-publish-skill - Source note for the JamesWuHK WeChat Channels / 视频号 Playwright publishing skill, local install, Windows build fix, smoke tests, and live-account safety boundaries.

• 2026-05-17-llm-hex-anime-video-project - Source note for the D-drive LLM ability-hexagon anime personification video project and its 视频号 publishing package.

• 2026-05-17-anbeime-frontend-design-and-chrome-automation - Source note for Anbeime frontend-design and chrome-automation skills, including local install paths and smoke-test results.

• 2026-05-17-larksuite-cli-feishu-bridge - Source note for official larksuite/cli, installed lark-* skills, local feishu-bridge, smoke tests, and OAuth-gated validation status.

• 2026-05-17-feishu-form-fill-session - Source note for the live Feishu shared Base form fill/submit session and browser-automation lessons.

• 2026-05-17-readme-blueprint-generator-skill - Source note for the GitHub awesome-copilot README specialist skill, local adaptation, install paths, and validation.

• 2026-05-17-email-assistant-skill - Source note for the customized email-assistant skill derived from Benchflow Gmail scripts, including API-first design, local auth paths, safety gates, and validation.

• 2026-05-17-communication-assistant-skill - Source note for the unified communication router skill, its adapter landscape, shared outbox, and prefill-first safety model.

• 2026-05-18-user-feedback-model-collaboration-context-intake - Chat source note for user feedback on DeepSeek delegation intent, partner context packs, and avoiding toy external-reference summaries.

• 2026-05-17-opencode-cc-pixelcat-setup - Source note for the D-drive OpenCode, Claude Code, and PixelCat setup, active D:\devtools paths, smoke-test results, and the 2026-05-18 upstream-credential outage diagnosis.

• 2026-05-17-research-project-roots-deep-review - Public-safe source note for the deep read-only review of analog-agent, Uncertainty/Pony-Rec, TRUCE-Rec, and TGL-Rec local roots.

• 2026-05-16-canvas-hack-recovery-note - User-provided Canvas outage/hack note; expected recovery 2026-05-18 and blocks syllabus retrieval for UMN transfer-credit review.

• 2026-w20-ai4s-protein - 2026-W20 AI4S protein/biology research pick.

• 2026-w20-ai4eda-analog - 2026-W20 AI4EDA analog circuit design research pick.

• 2026-w20-llm4rec - 2026-W20 LLM4Rec research pick.

• 2026-w20-llm - 2026-W20 LLM research pick.

• 2026-w20-ai - 2026-W20 AI research pick.

• 2026-05-16-frontend-frameworks-public-corpus - Batch ingest of frontend project framework GitHub repositories, release summaries, reuse metadata, language profiles, and candidate discovery.

• 2026-05-16-frontend-frameworks-public-corpus - Batch ingest of frontend framework GitHub repositories, release summaries, language profiles, and candidate discovery.

• 2026-05-16-yao-shunyu-public-corpora - Batch ingest of the ysymyth and alfredyao public corpora.

• 2026-05-16-qq-doc-pvz-fusion-mobile-pc - Tencent Docs source note for Plants vs. Zombies Fusion Edition mobile and PC; metadata visible, body permission-gated.

• 2026-05-16-lidang-public-corpus - Batch ingest of Lidang YouTube RSS, canonical X profile, auxiliary mirror probes, HTML snapshots, and dedupe/backfill rules.

• x-status-2041009459175026831-sp500-nasdaq100 - Manual evidence note for Lidang’s 买SP500/纳100 profile/status clue.

• 2026-05-16-lidang-public-corpus - Batch ingest of Lidang YouTube RSS, canonical X profile, and auxiliary mirror probes.

• 2026-05-15-karpathy-public-corpus - Batch ingest of Karpathy’s GitHub repositories, official homepage, blog RSS, YouTube feed, gists, and curated status links.

• 2026-05-15-openai-cookbook - Batch ingest of the official OpenAI Cookbook developers index and MIT-licensed GitHub source mirror.

• 2026-04-21-llm-wiki-pattern - Seed source note based on the user-provided idea brief describing the LLM wiki pattern.

• 2026-04-22-karpathy-llm-wiki-zh-compilation - User-provided Chinese compilation of the Karpathy article plus selected community commentary.

• 2026-04-22-security-course-self-study-guide - User-provided security course self-study guide describing AI rules, trustworthiness, and suggested week-one study behavior.

• 2026-04-22-5eid0-venus-project-course-materials - Local project-course materials for TU/e 5EID0, including the Venus manual, kickoff slides, and team sheet.

• 2026-04-21-nh-baseline-paper-set - External local paper set with 21 PDFs from the NH baseline folder under Uncertainty-LLM4Rec.

• 2026-04-21-nr-baseline-paper-set - External local paper set with 14 PDFs from the NR baseline folder under Uncertainty-LLM4Rec.

• 2026-04-21-recommendation-paper-library - Larger external local library of LLM-and-recommendation papers, with PDFs plus topic/project subfolders.

• 2026-05-08-university-of-minnesota-housing-contract-2026-2027 - University of Minnesota Housing & Residential Life contract terms, key dates, fees, cancellation/release rules, and resident obligations for 2026-2027.

• 2026-05-10-research-project-roots - Read-only survey of Vipin’s main local research project roots under D:/Research.

• 2026-05-10-d-drive-portfolio-and-undergraduate-roots - Public-safe content-nature inventory for D-drive portfolio, undergraduate project, and study roots.

• 2026-05-16-venus-team-project-github-archive - Migration note for the Venus Team 28 GitHub archive, including GitLab project snapshot, branch snapshots, UI module location, exclusions, and validation.

• 2026-05-15-qmq-app-us-visa-slot-service - User-provided and publicly checked pointer to QMQ App, a US visa CGI appointment slot service.

• 2026-05-15-d-drive-healthcare-and-skill-roots - Public-safe routing ingest for D:/Healthcare/Medora and the five D:/Skill repositories.

• 2026-05-16-skill-source-repository-trace - Upstream/origin trace for local D-drive skill repositories and their original public sources.

• 2026-05-15-terraria-local-save-files - Local Terraria save-file discovery, backup location, file-type summary, and inventory pointers.

• 2026-05-15-lu-kan-profile-chat - User-provided professional profile for 闄嗕緝 / Lu Kan across architecture, exhibition design, cultural tourism, rural revitalization, and AIGC workflows.

• qingtian-jay-chou-japanese-cover - Candidate song entry for the Japanese-cover 晴天 remembered with 小柔Channel.

• hare-no-hi-seiten - Song entry for 晴れの日（晴天）, including liked versions and playable-link TODOs.

Analyses

• 2026-06-04-weiping-wiki-rename-upgrade - Rename/upgrade record for the WEIPING_WIKI identity, agentmemory-first boundary, public/private separation, and research-integrity no-change scope.

• 2026-05-18-multi-agent-collaboration-architecture-review - Decision record for which Claude Code / DeepSeek collaboration improvements to adopt now, defer, or reject.

• html-ppt-template-workflow-results - Results and future guardrails from the first frontend-slides plus beautiful-html-templates deck-generation tests.

• weekly-research-digest-2026-w20 - 2026-W20 weekly AI, LLM, LLM4Rec, AI4EDA, and AI4S inspiration digest.

• frontend-framework-reuse-map - Reuse and migration map for captured frontend project frameworks.

• frontend-project-shell-taxonomy - Category map of captured frontend project frameworks by reusable shell role.

• public-corpus-ingest-workflow - Reusable workflow for complete public-person corpus ingests with automation and safe public indexing.

• shunyu-yao-paper-map - Paper, thesis, talk, and scholarly pointer map for the two Shunyu Yao corpora.

• shunyu-yao-project-taxonomy - Project and repository taxonomy across the two Shunyu Yao corpora.

• lidang-weekly-digests - Index of weekly digest batches and high-signal Lidang public idea pages.

• lidang-idea-taxonomy - Category map for Lidang public ideas across AI coding, career, immigration, startup, education, governance, and heuristics.

• karpathy-idea-map - Synthesis map of Karpathy’s reusable technical heuristics and public idea surfaces.

• karpathy-project-taxonomy - Category map of Karpathy public corpus items by implementation and idea domain.

• openai-cookbook-taxonomy - Category map of mirrored OpenAI Cookbook examples and articles.

• 2026-04-21-vipin-wiki-bootstrap - Initial implementation memo describing how this repository instantiates the LLM wiki idea.

• 2026-04-21-llm-rec-research-map - First-pass map of the LLM-for-recommendation literature represented in the current local paper collections.

• 2026-04-22-llm-wiki-skill-alignment - System alignment memo describing how this repository now matches the reference llm-wiki-skill operating model.

• 2026-05-10-vipin-research-project-map - Cross-project map of Vipin’s active research portfolio and claim boundaries.

• 2026-05-17-research-project-workbench-audit - Deep workbench audit for analog-agent plus the three local recommendation research projects, including tracked-file coverage, artifact inventory, claim gates, and collaboration-rule interpretation.

Synthesis

• synthesis-home - Entry page for the wiki/synthesis/ section.
• synthesis sessions - Session-level synthesis notes and crystallized checkpoints.
• 2026-04-22-karpathy-upgrade-session - Session note recording the upgrade from a toy-like bootstrap toward a reader-aware, catalog-backed knowledge operating system.

Timelines

• timelines-home - Entry page for the wiki/timelines/ section.

Queries

• queries-home - Entry page for durable saved answers.
• 2026-05-18-pixelcat-cc-502-credentials-disabled-runbook - Runbook for diagnosing and fixing PixelCat/ccmax HTTP 502 where cc works locally but all upstream credentials are disabled.
• 2026-05-18-vscode-black-screen-repair-runbook - Detailed Windows VS Code black-screen repair runbook with diagnosis, fixes, backups, and next-time fast path.
• 2026-05-17-where-can-local-cc-be-used - Short answer on where Vipin’s local cc/Claude Code setup can be used: terminal, Codex, VSCode/OpenCode, PixelCat-backed local tools, and why generic web/app chat is separate.
• 2026-05-17-how-to-ask-nuwa-to-distill-lidang - Copy-paste prompt for using Nuwa and Opus to update the Lidang perspective skill.
• 2026-04-22-how-to-reduce-github-submit-prompts - Workflow note on why GitHub updates may require approval prompts and how to reduce the friction.
• 2026-04-22-what-is-dbm - Short explanation of the dBm formula, its 1 mW reference, and common power conversion rules.
• 2026-04-22-what-is-spectroscopy - Short explanation of the term spectroscopy in a wave optics course context.
• 2026-04-22-how-are-k-omega-and-v-derived - Step-by-step derivation of k = 2pi/lambda, omega = 2pi f, and v = omega/k = f lambda.
• 2026-04-22-what-to-pay-attention-to-in-security-self-study-guide - Practical summary of what matters most in the security course self-study guide.
• 2026-04-22-what-to-pay-attention-to-in-5eid0-project-course - Practical summary of what matters most in the TU/e 5EID0 Venus project course.
• 2026-04-22-how-to-position-yourself-for-embedded-software-in-5eid0 - How to present and scope an embedded-software role in the 5EID0 team project.
• 2026-04-22-computer-software-vs-embedded-software-in-5eid0 - Comparison of computer software/UI versus embedded software roles for Vipin in the 5EID0 project.
• 2026-04-22-computer-software-ui-role-plan-for-5eid0 - Execution plan for Vipin’s chosen computer software/UI role in the 5EID0 project.
• 2026-04-22-what-language-for-5eid0-computer-software-ui - Recommendation to start the 5EID0 base-station UI in Python with MQTT support and simple map visualization.
• 2026-04-22-what-do-i-need-from-teammates-for-venus-basestation - Interface checklist describing what Vipin needs from teammates to build the base-station software in parallel.
• 2026-04-22-what-can-i-finish-independently-for-venus-basestation - Clear boundary between what Vipin can already complete in the basestation now and what still depends on teammates.
• 2026-04-23-what-causes-the-color-of-the-sky - Optics explanation that the sky’s color mainly comes from the electric field of light interacting with electrons and producing scattering.
• 2026-04-23-why-is-e-equals-cb - Explanation that E = cB in a vacuum electromagnetic wave follows from k 脳 E = 蠅B together with 蠅/k = c.
• 2026-04-23-what-does-codex-full-access-mean - Explanation of what broader Codex filesystem access means and the rule that project-external edits require explicit request or confirmation.
• 2026-04-23-do-e-and-b-contribute-equally-to-light-energy - Explanation that vacuum light energy is shared equally between the electric and magnetic field contributions.
• 2026-04-27-how-to-ask-teammates-for-basestation-interface-info - Reusable Chinese message template for asking teammates for MQTT topics, payload samples, coordinate definitions, and other basestation integration details.
• 2026-05-03-how-to-convert-this-spectrum-to-time-domain - Step-by-step explanation of how a spectrum with lines at -100 Hz, 0 Hz, and +100 Hz maps to a time-domain cosine plus DC offset.
• 2026-05-03-how-to-understand-flat-top-sampling-questions - Practical explanation of how to identify flat-top sampling and its main hold-effect amplitude distortion.
• 2026-05-03-what-is-an-adc-and-why-10-bit-means-1024-levels - Explanation of what an ADC is and why a 10-bit ADC has 1024 quantization levels.
• 2026-05-05-how-to-understand-quantization-and-bit-rate - Explanation of quantization levels, bits per sample, sampling frequency, bitrate, and bit period.
• 2026-05-05-why-pe-is-not-proportional-to-input-snr - Clarification that error probability depends on input SNR but normally decreases as input SNR increases.
• 2026-05-05-how-to-write-venus-project-problem-statement - Reusable English problem statement options for the 5EID0 Venus project group report.
• 2026-05-05-how-to-solve-oauth-rt-rbac-security-quiz - Reasoning guide for a security quiz on AAL, OAuth delegation, RT rules, replay attacks, and hierarchical RBAC.
• 2026-05-07-what-does-depicts-mean - Short vocabulary note explaining depicts as shows, describes, portrays, or represents.
• 2026-05-08-how-to-integrate-venus-basestation-with-team-gitlab - Git workflow plan for moving Vipin’s Venus basestation/UI work into the team GitLab repository on a dedicated branch.
• 2026-05-08-which-umn-meal-plan-to-choose - Practical recommendation for choosing between UMN Unlimited and Unlimited Plus residence hall meal plans.
• 2026-05-08-how-to-choose-umn-apartment-vs-residence-hall-for-private-bedroom - Practical UMN room-preference strategy for balancing low cost with a private sleeping room.
• 2026-05-16-what-indexes-did-lidang-recommend - Clarifies that Lidang’s SP500/纳100 phrase means S&P 500 and Nasdaq-100, not S&P 100.

Shunyu Yao Public Corpora

• Hubs
  • shunyu-yao-public-corpora
  • shunyu-yao-project-taxonomy
  • shunyu-yao-paper-map
• Items
  • alfredyao - alfredyao, canonical-homepage
  • Shunyu Yao CV - alfredyao, cv-pdf-metadata
  • alfredyao.github.io - alfredyao, github-repository
  • behavior_cloning_pytorch - alfredyao, github-repository
  • Numpy_classifier - alfredyao, github-repository
  • Pytorch_classifier - alfredyao, github-repository
  • Quantum_chaos_Diffusion_reaction - alfredyao, github-repository
  • SER_fine_tuning - alfredyao, github-repository
  • LAMIC: Layout-Aware Multi-Image Composition via Scalability of Multimodal Diffusion Transformer - alfredyao, paper
  • Notes on solvable models of many-body quantum chaos - alfredyao, paper
  • Physics3D: Learning Physical Properties of 3D Gaussians via Video Diffusion - alfredyao, paper
  • Quantum Energy Teleportation versus Information Teleportation - alfredyao, paper
  • CV - alfredyao, paper-or-pdf
  • post 2025 10 06 - alfredyao, post
  • post 2025 11 18 - alfredyao, post
  • ysymyth blog - ysymyth, blog-index
  • ysymyth - ysymyth, canonical-homepage
  • 3D-SDN - ysymyth, github-repository
  • awesome-language-agents - ysymyth, github-repository
  • ec-nl - ysymyth, github-repository
  • ReAct - ysymyth, github-repository
  • ysymyth.github.io - ysymyth, github-repository
  • reflexion - ysymyth, linked-github-repository
  • intercode - ysymyth, linked-github-repository
  • SWE-agent - ysymyth, linked-github-repository
  • SWE-bench - ysymyth, linked-github-repository
  • WebShop - ysymyth, linked-github-repository
  • tau-bench - ysymyth, linked-github-repository
  • tree-of-thought-llm - ysymyth, linked-github-repository
  • [[sources/shunyu-yao/ysymyth/paper-bench-a-benchmark-for-tool-agent-user-interaction-in-real-world-domains-b95f346da4|$\tau$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains]] - ysymyth, paper
  • Can Language Models Solve Olympiad Programming? - ysymyth, paper
  • CL-bench Life: Can Language Models Learn from Real-Life Context? - ysymyth, paper
  • CL-bench: A Benchmark for Context Learning - ysymyth, paper
  • Cognitive Architectures for Language Agents - ysymyth, paper
  • Contextual Experience Replay for Self-Improvement of Language Agents - ysymyth, paper
  • Controllable Text Generation for Large Language Models: A Survey - ysymyth, paper
  • HY-Embodied-0.5: Embodied Foundation Models for Real-World Agents - ysymyth, paper
  • InterCode: Standardizing and Benchmarking Interactive Coding with Execution Feedback - ysymyth, paper
  • Multi-objective Evolution of Heuristic Using Large Language Model - ysymyth, paper
  • OS-Copilot: Towards Generalist Computer Agents with Self-Improvement - ysymyth, paper
  • Prompting Large Language Models to Tackle the Full Software Development Lifecycle: A Case Study - ysymyth, paper
  • Provable Limitations of Acquiring Meaning from Ungrounded Form: What Will Future Language Models Understand? - ysymyth, paper
  • ReAct: Synergizing Reasoning and Acting in Language Models - ysymyth, paper
  • Reflexion: Language Agents with Verbal Reinforcement Learning - ysymyth, paper
  • SEAL: Self-Evolving Agentic Learning for Conversational Question Answering over Knowledge Graphs - ysymyth, paper
  • SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering - ysymyth, paper
  • SWE-bench: Can Language Models Resolve Real-World GitHub Issues? - ysymyth, paper
  • Tree of Thoughts: Deliberate Problem Solving with Large Language Models - ysymyth, paper
  • WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents - ysymyth, paper
  • When a language model is optimized for reasoning, does it still show embers of autoregression? An analysis of OpenAI o1 - ysymyth, paper
  • Computer-Using Agent system card - ysymyth, paper-or-pdf
  • paper Dissertation finalized - ysymyth, paper-or-pdf
  • paper fpo - ysymyth, paper-or-pdf
  • SWE-agent paper - ysymyth, paper-or-pdf
  • post benchmarking ai agents - ysymyth, post
  • Computer-Using Agent BibTeX - ysymyth, project-or-public-release
  • Computer-Using Agent blogpost - ysymyth, project-or-public-release
  • Computer-Using Agent product - ysymyth, project-or-public-release
  • Deep Research blogpost - ysymyth, project-or-public-release
  • Google AI blogpost - ysymyth, project-or-public-release
  • InterCode project - ysymyth, project-or-public-release
  • Quanta Magazine - ysymyth, project-or-public-release
  • ReAct project - ysymyth, project-or-public-release
  • SWE-agent project - ysymyth, project-or-public-release
  • SWE-bench project - ysymyth, project-or-public-release
  • WebShop demo - ysymyth, project-or-public-release
  • WebShop project - ysymyth, project-or-public-release
  • On Formulating and Evaluating Language Agents - ysymyth, talk-or-slides
  • Re-thinking Reinforcement Learning in the Era of Large Language Models - ysymyth, talk-or-slides
  • talk watch - ysymyth, talk-or-slides
  • 从语言模型到语言智能体 - ysymyth, talk-or-slides

Lidang Public Ideas Corpus

• Weekly digests
  • 2026-w20 digest
• High-signal single items
  • 立党直播022 深入解析manus事件、阿里林俊阳离职等时事热点
  • 一小时长聊：AI正在摧毁科技互联网行业，让程序员大批量失业？为什么程序员危机感空前强烈？为什么中国大学计算机专业教育质量太差？大学还有存在的意义吗？
  • 立党讲座27：我为什么推荐了整整10年的Georgia Tech online CS master（佐治亚理工在线CS硕士）？7000美元的在线硕士可以转码成功吗？有哪些优势和劣势？如何读好这类项目？

Karpathy Public Corpus

• Browser / JavaScript ML experiments
  • calorie
  • tsnejs
• LLM education / from-scratch pedagogy
  • cryptos
• LLM training and inference systems
  • build-nanogpt
  • cpython
  • examples
  • hn-time-capsule
  • karpathy.github.io
  • llama2.c
  • llm.c
  • LLM101n
  • minbpe
  • minGPT
  • nanochat
  • nanoGPT
  • ng-video-lecture
  • randomfun
  • reader3
  • rendergit
  • transformers
• Minimal implementations
  • An efficient, batched LSTM.
  • Batched L2 Normalization Layer for Torch nn package
  • Efficient LSTM cell in Torch
  • Git Commit Message AI
  • Google slides in present form shows the next slide, but it is tiny and very difficult to see. This CSS hacks it so that the next slide is large.
  • hacky stablediffusion code for generating videos
  • HELLO.md , written by Claude Opus 4.6 when asked to be free in a directory on my computer
  • llm-wiki
  • microgpt
  • numpy
  • Natural Evolution Strategies (NES) toy example that optimizes a quadratic function
  • pytorch strangeness
  • Training a Neural Network ATARI Pong agent with Policy Gradients from raw pixels
  • deep-vector-quantization
  • EigenLibSVM
  • find-birds
  • gitstats
  • karpathy
  • lifejs
  • MatlabWrapper
  • notpygamejs
  • optim
  • pytorch-made
  • pytorch-normalizing-flows
  • Random-Forest-Matlab
  • scholaroctopus
  • sqlitedict
  • twoolpy
• Neural network fundamentals
  • char-rnn
  • convnetjs
  • lecun1989-repro
  • makemore
  • micrograd
  • neuraltalk
  • neuraltalk2
  • nn-zero-to-hero
  • recurrentjs
• Personal heuristics / AI philosophy / learning advice
  • (started posting on Medium instead)
  • A from-scratch tour of Bitcoin in Python
  • A Recipe for Training Neural Networks
  • A Survival Guide to a PhD
  • Biohacking Lite
  • Deep Neural Nets: 33 years ago and 33 years from now
  • Deep Reinforcement Learning: Pong from Pixels
  • microgpt
  • Short Story on AI: A Cognitive Discontinuity.
  • Short Story on AI: Forward Pass
  • Curated Karpathy status 1013244313327681536
  • Curated Karpathy status 1018544870112165888
  • Curated Karpathy status 1031266180240093184
  • Curated Karpathy status 1076927879995285504
  • Curated Karpathy status 1080700888266637314
  • Curated Karpathy status 1084641729288884224
  • Curated Karpathy status 1106605823000666112
  • Curated Karpathy status 1127792584380706816
  • Curated Karpathy status 1133605485477634048
  • Curated Karpathy status 1158225988943089664
  • Curated Karpathy status 1206064174129930240
  • Curated Karpathy status 1231378194948706306
  • Curated Karpathy status 1236737502200791041
  • Curated Karpathy status 1274472300390150144
  • Curated Karpathy status 1279130562913951744
  • Curated Karpathy status 1282524334498541568
  • Curated Karpathy status 1284660899198820352
  • Curated Karpathy status 1292239731707666432
  • Curated Karpathy status 1297718122913308673
  • Curated Karpathy status 1311884485676294151
  • Curated Karpathy status 1320541466112843777
  • Curated Karpathy status 1325154823856033793
  • Curated Karpathy status 1328547710966743040
  • Curated Karpathy status 1331834386057105409
  • Curated Karpathy status 1338274248037597185
  • Curated Karpathy status 1342194661381836800
  • Curated Karpathy status 1350503355299205120
  • Curated Karpathy status 1358483726447988736
  • Curated Karpathy status 1365784221780828162
  • Curated Karpathy status 1389712138982158339
  • Curated Karpathy status 1398374159340634114
  • Curated Karpathy status 1401267972044328961
  • Curated Karpathy status 1429478697946402816
  • Curated Karpathy status 1435827240286109702
  • Curated Karpathy status 1439676583690715140
  • Curated Karpathy status 1442957114415063046
  • Curated Karpathy status 1467180759932801024
  • Curated Karpathy status 1468370605229547522
  • Curated Karpathy status 1470285570924417024
  • Curated Karpathy status 1486215976559398915
  • Curated Karpathy status 1497652099701620738
  • Curated Karpathy status 1504975036037406722
  • Curated Karpathy status 1513420766146084865
  • Curated Karpathy status 1515357035780616192
  • Curated Karpathy status 1515361631366266886
  • Curated Karpathy status 1528453604515778560
  • Curated Karpathy status 1532894698439774208
  • Curated Karpathy status 1550903910633680897
  • Curated Karpathy status 1558569874447671296
  • Curated Karpathy status 1558863432023126020
  • Curated Karpathy status 1564698516164730880
  • Curated Karpathy status 1574906895453675521
  • Curated Karpathy status 1582807367988654081
  • Curated Karpathy status 1590604672557252609
  • Curated Karpathy status 1593091486148489216
  • Curated Karpathy status 1593417987687473152
  • Curated Karpathy status 1599152286672248832
  • Curated Karpathy status 1599852921541128194
  • Curated Karpathy status 1600012576825360384
  • Curated Karpathy status 518489390050340864
  • Curated Karpathy status 543115142561804288
  • Curated Karpathy status 784212363062816768
  • Curated Karpathy status 801621764144971776
  • Curated Karpathy status 828022882215174144
  • Curated Karpathy status 868178954032513024
  • Curated Karpathy status 893576281375219712
  • Curated Karpathy status 904807765931024384
  • Curated Karpathy status 919396251849728001
  • Curated Karpathy status 961307010246492160
  • Curated Karpathy status 986454492534128640
  • Karpathy personal homepage
• Research automation / agentic science
  • autoresearch
  • jobs
  • llm-council
  • reinforcejs
  • scriptsbots
  • tf-agent
• Research tooling and paper workflows
  • arxiv-sanity-lite
  • arxiv-sanity-preserver
  • covid-sanity
  • nipspreview
  • paper-notes
  • researchlei
  • researchpooler
• Talks, courses, and videos
  • 1hr Talk Intro to Large Language Models
  • Building makemore Part 2: MLP
  • Building makemore Part 3: Activations & Gradients, BatchNorm
  • Building makemore Part 4: Becoming a Backprop Ninja
  • Building makemore Part 5: Building a WaveNet
  • Deep Dive into LLMs like ChatGPT
  • How I use LLMs
  • Let’s build GPT: from scratch, in code, spelled out.
  • Let’s build the GPT Tokenizer
  • Let’s reproduce GPT-2 (124M)
  • Stable diffusion dreams of psychedelic faces
  • Stable diffusion dreams of steampunk brains
  • Stable diffusion dreams of tattoos
  • The spelled-out intro to language modeling: building makemore
  • The spelled-out intro to neural networks and backpropagation: building micrograd
• Tokenization and language modeling
  • rustbpe
• Vision / multimodal / captioning
  • forestjs
  • nn
  • simple-amt
  • svmjs
  • ulogme

OpenAI Cookbook Mirror

• Agents SDK / agent workflows
  • Agent Improvement Loop
  • Agentkit Walkthrough
  • App Assistant Voice Agents
  • Building Consistent Workflows Codex Cli Agents Sdk
  • Building Reliable Agents Memory Compaction
  • Computer Use With Daytona
  • Context Personalization
  • Dispute Agent
  • Evaluate Agents
  • Multi Agent Portfolio Collaboration
  • Orchestrating Agents
  • Parallel Agents
  • Readme
  • Sandboxed Code Migration Agent
  • Secure Code Interpreter Tool For Llm Agents
  • Session Memory
• ChatGPT / GPT Actions
  • .Gpt Action Getting Started
  • Azure Ai Search With Azure Functions And Gpt Actions In Chatgpt
  • Chatgpt Agents Sales Meeting Prep
  • Chatgpt Prompt Guide
  • Getting Started With Bigquery Vector Search And Openai
  • Gpt Action Bigquery
  • Gpt Action Box
  • Gpt Action Canvas
  • Gpt Action Confluence
  • Gpt Action Github
  • Gpt Action Gmail
  • Gpt Action Google Calendar
  • Gpt Action Google Drive
  • Gpt Action Googleads Adzviser
  • Gpt Action Jira
  • Gpt Action Notion
  • Gpt Action Outlook
  • Gpt Action Pinecone Retool Rag
  • Gpt Action Redshift
  • Gpt Action Retool Workflow
  • Gpt Action Salesforce
  • Gpt Action Salesforce Gong
  • Gpt Action Sharepoint Doc
  • Gpt Action Sharepoint Text
  • Gpt Action Snowflake Direct
  • Gpt Action Snowflake Middleware
  • Gpt Action Sql Database
  • Gpt Action Trayai Apim
  • Gpt Action Workday
  • Gpt Action Zapier
  • Gpt Middleware Aws Function
  • Gpt Middleware Azure Function
  • Gpt Middleware Google Cloud Function
  • Logs Platform
• Codex / coding agents
  • Autofix Github Actions
  • Build Code Review With Codex Sdk
  • Build Iterative Repair Loops With Codex
  • Code Modernization
  • Codex Exec Plans
  • Codex Prompting Guide
  • Jira Github
  • Secure Quality Gitlab
  • Using Goals In Codex
• Deep Research / MCP
  • Databricks Mcp Cookbook
  • Mcp Tool Guide
• Evaluation / eval flywheels
  • Building Resilient Prompts Using An Evaluation Flywheel
  • Bulk Experimentation
  • Completion Monitoring
  • Custom Llm As A Judge
  • Elasticsearch Retrieval Augmented Generation
  • Evalsapi Audio Inputs
  • Evalsapi Image Inputs
  • Evaluate Rag With Llamaindex
  • Ft Retrieval Augmented Generation Qdrant
  • Getting Started With Openai Evals
  • How To Call Functions For Knowledge Retrieval
  • How To Eval Abstractive Summarization
  • How To Evaluate Llms For Sql Generation
  • Image Evals
  • Mcp Eval Notebook
  • Realtime Eval Guide
  • Receipt Inspection
  • Regression
  • Responses Evaluation
  • Selecting A Model Based On Stripe Conversion
  • Structured Outputs Evaluation
  • Tools Evaluation
  • Web Search Evaluation
• Fine-tuning / reinforcement fine-tuning
  • Chat Finetuning Data Prep
  • Fine Tune Korean
  • Fine Tune Transfomers
  • Fine Tuning Direct Preference Optimization Guide
  • Fine Tuning For Function Calling
  • Gpt Finetuning With Wandb
  • How To Finetune Chat Models
  • Reinforcement Fine Tuning
  • Reinforcement Finetuning Healthbench
  • Vision Fine Tuning On Gpt4o For Visual Question Answering
• General OpenAI API patterns
  • Assistants Api Overview Python
  • Batch Processing
  • Building W Rt Mini
  • Completions Usage Api
  • Data Extraction Transformation
  • Developing Hallucination Guardrails
  • Enhance Your Prompts With Meta Prompting
  • Gpt4 1 Prompting Guide
  • How To Build An Agent With The Node Sdk
  • How To Call Functions With Chat Models
  • How To Handle Rate Limits
  • How To Use Guardrails
  • How To Use Moderation
  • Introduction To Gpt4o
  • Named Entity Recognition To Enrich Text
  • Optimize Prompts
  • Prompt Caching 201
  • Prompt Caching101
  • Prompt Migration Guide
  • Reproducible Outputs With The Seed Parameter
  • Sdg1
  • Skills In Api
  • Summarizing Long Documents
  • Unit Test Writing Using A Multi Step Prompt With Older Completions Api
  • Using Logprobs
  • What Makes Documentation Good
• GPT-5 / reasoning / prompting
  • Build A Coding Agent With Gpt 5.1
  • Gpt 5 1 Prompting Guide
  • Gpt 5 2 Prompting Guide
  • Gpt 5 Frontend
  • Gpt 5 New Params And Tools
  • Gpt 5 Prompting Guide
  • Gpt 5 Troubleshooting Guide
  • O3o4 Mini Prompting Guide
  • Prompt Optimization Cookbook
  • Prompt Personalities
  • Reasoning Function Calls
  • Using Reasoning For Data Validation
  • Using Reasoning For Routine Generation
• gpt-oss / open-weight deployment
  • Build Your Own Fact Checker Cerebras
  • Gpt Oss Safeguard Guide
  • Handle Raw Cot
  • Openai Harmony
  • Run Colab
  • Run Locally Lmstudio
  • Run Locally Ollama
  • Run Nvidia
  • Run Transformers
  • Run Vllm
  • Verifying Implementations
• Multimodal / image / video
  • Creating Slides With Assistants Api And Dall E3
  • Document And Multimodal Understanding Tips
  • Generate Images With Gpt Image
  • Generate Images With High Input Fidelity
  • Gpt With Vision For Video Understanding
  • Grounded Spatial Reasoning Layouts
  • How To Create Dynamic Masks With Dall E And Segment Anything
  • Image Gen 1.5 Prompting Guide
  • Image Gen Models Prompting Guide
  • Sora2 Prompting Guide
  • Tag Caption Images With Gpt4v
  • Using Gpt4 Vision With Function Calling
  • What Is New With Dalle 3
• RAG / retrieval / vector databases
  • Custom Image Embedding Search
  • Deeplake Langchain Qa
  • Elasticsearch Semantic Search
  • Embedding Wikipedia Articles For Search
  • File Search Responses
  • Generative Search With Weaviate And Openai
  • Getting Started With Azure Ai Search And Openai
  • Getting Started With Hologres And Openai
  • Getting Started With Polardb And Openai
  • Getting Started With Tair And Openai
  • How To Automate S3 Storage With Functions
  • How To Combine Gpt4o With Rag Outfit Assistant
  • Hyde With Chroma And Openai
  • Image Understanding With Rag
  • Introduction To Deep Research Api
  • Introduction To Deep Research Api Agents
  • Leveraging Model Distillation To Fine Tune A Model
  • Neon Postgres Vector Search Pgvector
  • Openai Wikipedia Semantic Search
  • Parse Pdf Docs For Rag
  • Philosophical Quotes Cassio
  • Philosophical Quotes Cql
  • Qa With Langchain Tair And Openai
  • Question Answering Using A Search Api
  • Rag With Graph Db
  • Readme
  • Readme
  • Readme
  • Readme
  • Readme
  • Readme
  • Readme
  • Readme
  • Search Reranking With Cross Encoders
  • Semantic Search
  • Semantic Search Using Mongodb Atlas Vector Search
  • Using Chroma For Embeddings Search
  • Using Myscale For Embeddings Search
  • Using Pinecone For Embeddings Search
  • Using Qdrant For Embeddings Search
  • Using Redis For Embeddings Search
  • Using Typesense For Embeddings Search
  • Using Vision Modality For Rag With Pinecone
  • Using Weaviate For Embeddings Search
  • Visualizing Embeddings In Kangas
  • Web Search With Google Api Bring Your Own Browser Tool
• Realtime / voice / transcription
  • Context Summarization With Realtime Api
  • Data Intensive Realtime Apps
  • Mcp Powered Agents Cookbook
  • One Way Translation Using Realtime Api
  • Realtime Out Of Band Transcription
  • Realtime Prompting Guide
  • Realtime Translation Guide
  • Running Realtime Api Speech On Esp32 Arduino Edge Runtime Elatoai
  • Speech Transcription Methods
  • Steering Tts
  • Voice Translation Into Different Languages Using Gpt 4O
  • Whisper Correct Misspelling
  • Whisper Processing Guide
  • Whisper Prompting Guide
• Responses API / tool orchestration
  • Reasoning Items
  • Responses Api Tool Orchestration
  • Responses Example
• Structured outputs / function calling
  • Function Calling Finding Nearby Places
  • Function Calling With An Openapi Spec
  • Structured Outputs Intro
  • Structured Outputs Multi Agent
  • Using Chained Calls For O1 Structured Outputs
  • Using Tool Required For Customer Service
• Third-party integrations
  • Agentic Governance Cookbook
  • Autonomous Agent Retraining
  • Chat With Your Own Data
  • Code Quality And Security Scan With Github Actions
  • Financial Document Analysis With Llamaindex
  • Functions
  • Model Selection Guide
  • Openai Monitoring With Wandb Weave
  • Temporal Agents

Topics

• frontend-frameworks-public - Curated public corpus for reusable frontend project shells, documentation/wiki frameworks, app frameworks, build tooling, and release ideas.
• shunyu-yao-public-corpora - Hub that keeps the two Shunyu Yao public corpora separate and searchable.
• pvz-fusion-edition - Routing topic for Plants vs. Zombies Fusion Edition mobile and PC setup sources.
• lidang-public-ideas - Hub for high-frequency Lidang public ideas with weekly digests, taxonomy, and source-confidence boundaries.
• karpathy-public-work - Hub for Karpathy’s public projects, posts, videos, gists, and curated idea links.
• openai-cookbook - Mirrored OpenAI Cookbook implementation library with categorized examples, articles, source hashes, and weekly refresh workflow.
• topics-home - Entry page for durable subject clusters.
• personal-knowledge-systems - Topic map for long-lived systems that accumulate, structure, and revisit knowledge.
• personal-music-preferences - Personal song memory, including liked versions and future follow-up notes.
• university-housing - Topic map for university housing deadlines, fees, cancellation/release rules, and contract obligations.
• research-projects - Topic map of Vipin’s active research projects and cross-cutting evidence rules.
• local-project-roots - Topic map for local D-drive project/archive roots, discovery rules, and safety boundaries.
• us-visa-appointments - Practical notes and safety boundaries for US visa appointment scheduling and slot-monitoring tools.
• healthcare-projects - Local healthcare project routing and public/private safety boundaries.
• agent-skill-repositories - Local skill repository routing for skill stores, generators, optimizers, engineering skill sets, project-local paper-writing skills, and publishing skills.
• terraria-save-archive - Local Terraria save archive, backup path, restore notes, and sorting heuristics.

Comparisons

• comparisons-home - Entry page for side-by-side evaluations and tradeoff notes.
• llm-wiki-vs-rag - Comparison between persistent wiki compilation and query-time retrieval.

Frontend Framework Corpus Generated Pages

Framework Entities

• Angular - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Astro - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Docusaurus - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Fumadocs - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Next.js - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Nextra - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Nitro - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Nuxt - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Quartz - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Qwik - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• React - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• React Router - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Solid - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Starlight - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Svelte - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• SvelteKit - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• TanStack Start - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Vite - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• VitePress - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.
• Vue - Project framework entity page with reuse profile, shell score, curated repos, releases, docs, and navigation.

Repository Sources

• angular - GitHub repo source page for Angular, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• angular-cli - GitHub repo source page for Angular, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• components - GitHub repo source page for Angular, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• docusaurus - GitHub repo source page for Docusaurus, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• react - GitHub repo source page for React, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• fumadocs - GitHub repo source page for Fumadocs, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• quartz - GitHub repo source page for Quartz, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• nitro - GitHub repo source page for Nitro, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• content - GitHub repo source page for Nuxt, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• image - GitHub repo source page for Nuxt, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• nuxt - GitHub repo source page for Nuxt, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• ui - GitHub repo source page for Nuxt, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• qwik - GitHub repo source page for Qwik, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• react.dev - GitHub repo source page for React, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• react-router - GitHub repo source page for Remix / React Router, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• remix - GitHub repo source page for Remix / React Router, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• nextra - GitHub repo source page for Nextra, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• solid - GitHub repo source page for Solid, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• solid-docs - GitHub repo source page for Solid, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• solid-start - GitHub repo source page for Solid, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• kit - GitHub repo source page for SvelteKit, Svelte, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• svelte - GitHub repo source page for Svelte, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• svelte.dev - GitHub repo source page for Svelte, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• router - GitHub repo source page for TanStack Start, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• next.js - GitHub repo source page for Next.js, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• vite - GitHub repo source page for Vite, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• vite-plugin-react - GitHub repo source page for Vite, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• core - GitHub repo source page for Vue, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• devtools - GitHub repo source page for Vue, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• docs - GitHub repo source page for Vue, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• pinia - GitHub repo source page for Vue, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• router - GitHub repo source page for Vue, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• vitepress - GitHub repo source page for VitePress, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• astro - GitHub repo source page for Astro, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• docs - GitHub repo source page for Astro, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.
• starlight - GitHub repo source page for Starlight, Astro, reuse profile, docs/demo links, language profile, release pointer, hashes, and metadata.

Release Sources

• angular v20.3.21 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• angular v21.2.13 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• angular v22.0.0-rc.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• angular-cli v20.3.26 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• angular-cli v21.2.11 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• angular-cli v22.0.0-rc.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• components v21.2.11 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• components v22.0.0-next.8 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• components v22.0.0-rc.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• docusaurus v3.10.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• docusaurus v3.10.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• docusaurus v3.9.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• react v19.0.6 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• react v19.1.7 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• react v19.2.6 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• local-md@0.2.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• sanity@0.0.4 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• fumadocs fumadocs-mdx@15.0.5 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• quartz v3.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• quartz v3.3 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• quartz v4.0.8 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• nitro v2.13.4 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• nitro v3.0.260415-beta - Captured GitHub release summary, project-shell idea bullets, and body hash.
• nitro v3.0.260429-beta - Captured GitHub release summary, project-shell idea bullets, and body hash.
• content v3.11.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• content v3.12.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• content v3.13.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• image v1.11.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• image v2.0.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• image v2.0.0-alpha.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• nuxt v3.21.5 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• nuxt v4.4.3 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• nuxt v4.4.5 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• ui v4.6.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• ui v4.7.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• ui v4.7.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• optimizer@2.1.0-beta.4 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• react@2.0.0-beta.35 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• router@2.0.0-beta.35 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• react-router react-router@7.14.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• react-router react-router@7.15.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• react-router react-router@7.15.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• remix static-middleware@0.4.8 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• remix test@0.3.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• remix ui@0.1.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• nextra nextra@4.6.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• nextra nextra-theme-blog@4.6.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• nextra nextra-theme-docs@4.6.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• solid v1.8.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• solid v1.9.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• solid v2.0.0-beta.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• start@1.3.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• start@1.3.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• start@1.3.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• kit@2.59.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• kit@2.60.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• kit@2.60.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• svelte svelte@5.55.5 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• svelte svelte@5.55.6 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• svelte svelte@5.55.7 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• start-plugin-core@1.170.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• vue-start@1.168.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• router release-2026-05-16-0506 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• next.js v16.3.0-canary.19 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• next.js v16.3.0-canary.20 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• next.js v16.3.0-canary.21 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• vite plugin-legacy@8.0.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• vite v8.0.12 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• vite v8.0.13 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• vite-plugin-react plugin-react@6.0.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• vite-plugin-react plugin-react-swc@4.3.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• vite-plugin-react plugin-rsc@0.5.26 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• core v3.5.34 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• core v3.6.0-beta.11 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• core v3.6.0-beta.12 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• devtools v8.1.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• devtools v8.1.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• devtools v8.1.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• pinia v3.0.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• pinia v3.0.3 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• pinia v3.0.4 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• router v5.0.5 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• router v5.0.6 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• router v5.0.7 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• vitepress v2.0.0-alpha.15 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• vitepress v2.0.0-alpha.16 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• vitepress v2.0.0-alpha.17 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• mdx@5.0.6 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• astro astro@6.3.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• astro astro@6.3.3 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• starlight@0.39.0 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• starlight@0.39.1 - Captured GitHub release summary, project-shell idea bullets, and body hash.
• starlight@0.39.2 - Captured GitHub release summary, project-shell idea bullets, and body hash.

Weekly Research Digests

• weekly-research-digests - Weekly AI/LLM/LLM4Rec/AI4EDA/AI4S inspiration picks.

Codex Prompt Corpus

• codex-prompt-corpus - High-quality selected local Codex prompts and automation prompts for reuse.
• codex-prompt-taxonomy - Category map and filtering rules for the Codex prompt corpus.

Codex Prompt Sources

• 这不是 BLOCKER，不要继续在 max_requests 这种 trivial safety mismatch 上停住。 - codex-user-prompt:7832b29cb044ac0c
• 你先做：接下来要做的是： - codex-user-prompt:7c237abb8652745e
• official_external_baselines.yaml and any provenance-oriented… - codex-user-prompt:a8d14c84ce2411d6
• 我建议你采用多agent工作，整理一下我们所有的文件和后续要走的清单和我们的计划，尽量详细一点。readme也重新更新一下。把 official base… - codex-user-prompt:66468b8bf1baafa5
• Please re-review the current fixed diff for the findings you raised. Focus on… - codex-user-prompt:1e996ee21e493a71
• repro checker。请只读检查当前计划文档改动是否满足 AGENTS.md 的复杂任务要求：证据路… - codex-user-prompt:320a3cc1dea995f9
• 以后都要记得自动commit push。然后我建议你不要那么保守，请你直接落成完整的prompt wiki，具体参考的六个项目，你要自己去看一下，学习他们… - codex-user-prompt:5eb688f439065259
• In D:\Research\Uncertainty, act as server-diagnostic planner for SETRec. Base… - codex-user-prompt:42ab3d53a4f1cb0e
• Read the docs folder first, especially: - codex-user-prompt:1687034557a9eb9a
• commit and push - codex-user-prompt:ca641720d17374f1
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:7cbf27b02da1e40b
• Automation: Light Crawl Lidang Public Ideas - codex-user-prompt:adf1736a425fa5cc
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:9d27f4d021416dd6
• Run the daily light capture for the Lidang public ideas corpus in the vipin w… - codex-automation-prompt:light-crawl-lidang-public-ideas
• Research Quality Monito… - codex-automation-prompt:reviewer-truce-rec
• 我批准审核 - codex-automation-prompt:worker1-truce-rec
• R2-real-LLM subgate returned PASS WITH MINOR FIXES. - codex-user-prompt:1add83f8a5d5e350
• 更新到github main， - codex-user-prompt:a49c2f62f5e1d0e9
• Phase 9C main accuracy multi-seed aggregation is complete. - codex-user-prompt:74ef89b3b1e3a340
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:ba3d86b5fa9d6b8f
• 你是实验复现和benchmark setting explorer。请在当前仓库 C:\Users\admin.codex\worktrees\da20… - codex-user-prompt:7a1ebba005a21194
• 方法 explorer。请在当前仓库 C:\Users\admin.codex\worktrees\da20\Uncertain… - codex-user-prompt:45801b82a3740170
• Files mentioned by the user: - codex-user-prompt:ef80b41b1c352a51
• 工作区：D:\Research\DoneBench。你不是独自在代码库中工作，请不要回退他人的改动。当前项目是 DoneBench benchmark M… - codex-user-prompt:c384604dfd126b3e
• 实现 AI-assisted audit 管线… - codex-user-prompt:5cd68ebeac91ce41
• 工作区 D:\Research\DoneBench。你不是独自在代码库中工作，不要回退他人改动。任务：补 Reproducibility package。… - codex-user-prompt:637b0181007b6d9d
• You own README.md plus the high-level planning docs: OFFICIAL_EXTERNAL_BASELI… - codex-user-prompt:befad3783ccc0580
• 这是我后面做的一些事：那我们做到了保证公平性的情况下其他都要官方原装实现的这个要求了吗 - codex-user-prompt:1d14a4a12207706b
• Read-only task for TGL-Rec. Audit current repo for runnable server workflow g… - codex-user-prompt:6ce8c3f5f372da8d
• Repository: D:\Research\TGL-Rec. Narrow audit task while parent edits: inspec… - codex-user-prompt:a127153fd0eaaa7a
• 在Security helper这个文件夹下，直接做一个网页项目，这个网页的要求很简单，用户能够在框内输入学号，然后直接产生对应lab的token，目前版… - codex-user-prompt:ddb6f72fd5a2895a
• 这样就可以了吧，其他还需要做什么，你能不能直接看看他们在gitlab其他模块已经做的东西，然后看看我们的computer software 和ui的ven… - codex-user-prompt:e5563d88490b1fce
• Files mentioned by the user: - codex-user-prompt:8c4152181895b3cc
• 请只读检查 D:\Research\Uncertainty 之外服务器官方 repo 无法本地访问时的替代：在本地仓库中查 main_prepare_ll… - codex-user-prompt:7ea12ee937162c4c
• 能不能让agent以后都记住，我们的大方向以及学长的一些话，精确，详细（你需要汲取我先前说的话和你的对应文档，做完一部分也要记得更新，不然一直陈旧的后面的… - codex-user-prompt:968add0df371b58b
• 采用多agent协作，推进下一步，同时梳理全部的项目和文档，整理适配于agent工作的文档，让之后的codex能够清楚的知道已经做了什么目前要做什么后面要… - codex-user-prompt:f4478b666788fc5b
• human audit 工作流现状。请重点阅读 don… - codex-user-prompt:b5307a63b3c9f880
• 做这些：跑 GPT-5.5 targeted audit，处理 ai_adjudication_queue_nonempty。 - codex-user-prompt:94869b67eddc78fd
• GPT-5.5 targeted audit worker D。只读审计 DoneBench 的这些任务：sheet_db_025, s… - codex-user-prompt:bae6b6b52b1920ce
• GPT-5.5 targeted audit worker C。只读审计 DoneBench 的这些任务：file_doc_024, f… - codex-user-prompt:e2d076c9c57d75b6
• 你觉得有必要用gpt5.2,跑full域吗。这个也请你来代劳一下，你可以搞多个小agent然后跑full域，最好就是如果中断了也能保留已经跑的部分，不至于… - codex-user-prompt:743f0f7a953b714a
• 重新生成这 100 个 human-audit queue 任务的 initial state、reference trace、final sta… - codex-user-prompt:0e8857057ed04891
• server docs for what a future… - codex-user-prompt:45b13bfd906afcbc
• Repository: D:\Research\TGL-Rec. Audit the docs and root files that future Co… - codex-user-prompt:29fc7586fdfc7208
• You are implementation-planner worker for TGL-Rec Phase 10. You are not alone… - codex-user-prompt:79dc7b01e932c221
• Review TRUCE-Rec’s current docs and method implementation for the user’s late… - codex-user-prompt:377d219b8fa7193f
• four_do… - codex-user-prompt:814438af7b7bd76e
• 多agent协作先做好这些，除非确实做不了，把 paper submission-ready 化 - codex-user-prompt:14c4c364bc502e0a
• tests aroun… - codex-user-prompt:b01b034ae755071e
• Implementation planner for TGL-Rec method upgrade. You are not alone in the c… - codex-user-prompt:a231a4b4e69ed34d
• server_runboo… - codex-user-prompt:35966d55a80d1da6
• M5, LLM-ESR official external baseline upgrade. Please inspect t… - codex-user-prompt:d8a788278565e797
• 查阅近年顶会或强相关 benchmark 项目，聚焦 realistic agent benchmarks：WebAr… - codex-user-prompt:34cd5eadf15df022
• reproducibility 型顶会项目：SWE-bench、SWE-benc… - codex-user-prompt:39e440060264ca1f
• Files mentioned by the user: - codex-user-prompt:ef245c1c14e4f2e4
• 我想利用harness engineering 的全套思路，重新构建这个项目，内容需要保留，但是架构需要做调整，并增加网页，这个项目需要一个网页，同时请你… - codex-user-prompt:44e6a80871db8715
• 我想利用harness engineering 的全套思路（自己去搜寻里面包含什么概念），重新构建这个项目，内容idea只做部分保留，给你大的权限删去已有… - codex-user-prompt:9834d6338190d1be
• 本地大重构与服务器就绪计划 - codex-user-prompt:c382ed7dfc607cee
• [[sources/codex-prompts/codex-user-prompt-986784ea9231d53e|https://github.com/appleweiping/analog-agent.git这个是目前的项目在github的状态，我想对codex说对…]] - codex-user-prompt:986784ea9231d53e
• 只读探索任务，不要修改任何文件。工作目录是 D:\Research\Agent-AI4EDA，真正仓库很可能在 D:\Research\Agent-AI4… - codex-user-prompt:df4060f2a683cce9
• model-based plannin… - codex-user-prompt:ea89b0a29355e38a
• 只读探索任务，不要修改文件。请把当前 analog-agent 与外部参考（AnalogGym、AnalogCoder、AnalogGenie、AICir… - codex-user-prompt:39cf3e1ec375383a
• Repository: D:\Research\UncertaintyProtein-AI4S. Plan-mode only. Do not edit… - codex-user-prompt:d97cc039572e7ad3
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:3fba5dbbca4ddf63
• Repository: D:\Research\UncertaintyProtein-AI4S. You are not alone in the cod… - codex-user-prompt:004a1e12ed7135e8
• Review branch for a coding task. Repo: D:\Research\Agent-AI4EDA\analog-agent,… - codex-user-prompt:0e4e86bce96e6794
• vector store +… - codex-user-prompt:dcbeb13718668c40
• 验证 agent”。仓库在 D:\Research\Research Recon Pro… - codex-user-prompt:39876879fb82e663
• 你是 Research Recon Prototype 的“故障修复 agent”。仓库在 D:\Research\Research Recon Prot… - codex-user-prompt:c7250a678fb7ec2e
• 你是 Research Recon Prototype 的真实用户模拟 Agent A。请在仓库 D:\Research\Research Recon P… - codex-user-prompt:2db804d0607875d2
• 你是 Research Recon Prototype 的真实用户模拟 Agent B。请在仓库 D:\Research\Research Recon P… - codex-user-prompt:2d49732a289386e2
• 目前这个项目，有一个问题就是按一下抓取，如果已经有20个了，他就一直都是20个。然后也不能拿掉一些论文，也许我的判断也不一定对，但是给我的直观感受是这样的… - codex-user-prompt:8af3a913c408de2f
• 做完这些，派六个agent，其中两个充当真实用户reviewer（不单单只是看一下界面那么简单，要实际使用（所有环节），多环境模拟，多api调用模拟，co… - codex-user-prompt:7c848f951a38aca7
• HF 项目调研，用来发现 Research Recon 应该补的 idea 和功能。… - codex-user-prompt:10c92f30b6c7753f
• top-conference 方向来扩展产品能力。只读，不修改 repo。请用网… - codex-user-prompt:e7deef72abf42369
• 你是 Implementation Agent 5，但先不要改代码。工作目录 D:\Research\Research Recon Prototype。等… - codex-user-prompt:5600cbbe57749c3b
• docs，避免碰 frontend dashboard 以免和主 agent 冲突。你不是一个人在代… - codex-user-prompt:0dbde13815f47abb
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:2adf219cf187bdbf
• In D:\Research\TRUCE-Rec, inspect the repository docs relevant to creating a… - codex-user-prompt:7579b6069a9d4ffc
• Prompt Wiki reference Agent 1: Study Langfuse as a reference for prompt manag… - codex-user-prompt:7cb930cc905d6bc4
• Prompt Wiki reference Agent 2: Study PromptLayer as a reference for prompt re… - codex-user-prompt:59d500a3a583948d
• Prompt Wiki reference Agent 3: Study Humanloop as a reference for prompt engi… - codex-user-prompt:b5eb230936e58acb
• Prompt Wiki reference Agent 4: Study promptfoo as a reference for prompt eval… - codex-user-prompt:2a5c1cfc8c981315
• community promp… - codex-user-prompt:7a8ca5729a329e44
• Agent 1: Read-only inspect D:\Research\vipin’s knowledgebase and D:\Research… - codex-user-prompt:b4f4dd5c74c33882
• Agent 3: Read-only inspect Prompt Wiki packages. Focus on preserving the prom… - codex-user-prompt:59c0990bf7c2d1a6
• Agent 4: Read-only inspect current GitHub Pages setup and wiki output in Prom… - codex-user-prompt:3955af5786171acd
• Read-only planning task. Do not modify files. Based on the user’s request, fo… - codex-user-prompt:7e30f2ebd3fa501e
• 你在电脑上应该能够看到两个vscode一个是本地的一个是我想去连接服务器，但我不知道为什么现在这么慢跑的，还有我也想问你，我该怎么样才能让codex在服务… - codex-user-prompt:adfcb36ff0fb8019
• In D:\Research\Uncertainty, inspect the existing official runner patterns and… - codex-user-prompt:2dae1affb72672ac
• [[sources/codex-prompts/codex-user-prompt-694c7787265c710d|你现在接手这个项目：chronicles-of-many-ages。]] - codex-user-prompt:694c7787265c710d
• 只做只读诊断，不要编辑文件。你是 Game Director Agent。仓库路径：D:\Game_develop\Chronicles of Many… - codex-user-prompt:fbbbe8ecc8565b19
• 只做只读诊断，不要编辑文件。你是 Simulation Architect Agent。仓库路径：D:\Game_develop\Chronicles o… - codex-user-prompt:6a7abc41c1a6c149
• Game Feel Agent。仓库路径：D:\Game_develop\Chronicles of Many… - codex-user-prompt:12d1a896754181e7
• 只做只读诊断，不要编辑文件。你是 Narrative Systems Agent。仓库路径：D:\Game_develop\Chronicles of M… - codex-user-prompt:5b325301863f416d
• 只做只读诊断，不要编辑文件。你是 Technical Refactor Agent。仓库路径：D:\Game_develop\Chronicles of… - codex-user-prompt:6342d2b82b9a3db9
• Integration Agent。仓库路径：D:\Game_develop\Chr… - codex-user-prompt:5a2cea62be6b31e1
• 我们要创建一个名为 Agent gamedevelopmentstudio 的项目。 - codex-user-prompt:902ac6d571a1fe84
• 请重写并强化根目录的 AGENTS.md。 - codex-user-prompt:8ef9777c4280712f
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:bb760ac5f9048f36
• 现在基于根目录 AGENTS.md，创建该项目的的制度层文档和模板。 - codex-user-prompt:9acd6933fba67cbb
• ，实现 Agent GD Studio 的最小客户旅程原型。 - codex-user-prompt:69c5c5deea528a19
• [[sources/codex-prompts/codex-user-prompt-d49b2d7b6fe89bb1|https://github.com/appleweiping/game-development-studio-agent.git帮我连接，分批次comm…]] - codex-user-prompt:d49b2d7b6fe89bb1
• Selected text: - codex-user-prompt:19dffc3d3fe29469
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:9ae29678b2e96198
• [[sources/codex-prompts/codex-user-prompt-9bfeb51892401a7f|你先连接一下这个仓库。https://github.com/appleweiping/Medora.git]] - codex-user-prompt:9bfeb51892401a7f
• Read the docs folder first. - codex-user-prompt:0c740691f12d242e
• Read the docs folder first. - codex-user-prompt:b5161ded533ee0f2
• Read the docs folder first. - codex-user-prompt:df18e1ea73b73f60
• Read the docs folder first. - codex-user-prompt:917ddae462b16361
• Read the docs folder first, especially: - codex-user-prompt:1ce5f3a9c1f44018
• 这是我跟gpt的对话：只想同步作者更新的话，也建议 fork + clone，但你尽量不要在自己的 fork 里改内容。这样同步最省心。 - codex-user-prompt:165dd7a3730aa431
• Context from my IDE setup: - codex-user-prompt:29fe98ad78455319
• Context from my IDE setup: - codex-user-prompt:3f7b781414e5edb2
• Context from my IDE setup: - codex-user-prompt:45b12b7f5fbaf1ab
• [[sources/codex-prompts/codex-user-prompt-9c6aed1261d62e88|https://github.com/appleweiping/skill.git这个是我fork了别人的skill收录项目，里面有一个video-cre…]] - codex-user-prompt:9c6aed1261d62e88
• Files mentioned by the user: - codex-user-prompt:97a5e41572b6ddb4
• 你负责为“professional concert pianist grand piano performing Canon in D elegant c… - codex-user-prompt:42a176a5e33670f5
• 我的意思是这样，我现在还想创建一些新的世界和人物的存档，我觉得我们可以搜寻一下github其他人的仓库或者资料来源，最好是那种高端的，热门的那种，跟我们不… - codex-user-prompt:e7bde27dcc7fecf0
• Karpathy他的完整github和他的所有项目和他的所有公开发布信息，这个你要仔细ingest，这个对我过分重要，主要就是得分好类，然后具体说说讲了什… - codex-user-prompt:4816ab78d91d8f6a
• 姚顺宇和姚顺雨这两个人的他们的完整github和他的所有项目和他的所有公开发布信息以及他们的所有论文，这个你要仔细ingest，这个对我过分重要，主要就是… - codex-user-prompt:7d055b980af160f8
• 热门前端框架，的完整github和他们的所有项目和他的所有公开发布信息，这个你要仔细ingest，这个对我过分重要，主要就是得分好类，然后具体说说讲了什么… - codex-user-prompt:4738c8ab2f376b85
• e-shadow-v6$ pgrep -af ‘run_week8_large_scale_10k_100neg|main_generate_llmesr… - codex-user-prompt:f68c64f7705aaa5d
• 我想生成一个真实感怀旧风格的视频，画面像第一人称视角走在中国杭州的街道上。整体氛围是温柔、安静、带一点青春回忆感，不要赛博朋克，不要旅游宣传片感，要像普通… - codex-user-prompt:c2622e24d0ebd143
• templated - codex-user-prompt:1770ed9eefb4da1c
• 有个问题，目前的这个前端他其实就是想把什么都展示在一个页面上了，他没有比方说有一侧是目录栏，然后我点击一个链接，然后主页面就跳转到一个相应的页面里，这个就… - codex-user-prompt:8db19ee20e4f667a
• 你负责分析用户给的两个本地参考视频的视觉风格，用于创建“钢琴家在音乐厅演奏 Canon in D”的 AI 视频生成项目。参考视频路径： - codex-user-prompt:0fd7b37b6f931eca
• 这是一个 vanilla 1.4.x large master getfixedboi 种子世界，作为非 toy 原创项目的 base world；我没有… - codex-user-prompt:bccbe8e206ebd388
• Read the docs folder first. - codex-user-prompt:15cf68518090bc02
• GPT-5.5 targeted audit worker A。只读审计 DoneBench 的这些任务：calendar_028, c… - codex-user-prompt:ac95e31f14dbed9f
• GPT-5.5 targeted audit worker B。只读审计 DoneBench 的这些任务：crm_workflow_03… - codex-user-prompt:659ddaeb49f6e481
• You are auditing DoneBench task quality for one domain. Workspace: D:\Researc… - codex-user-prompt:1fc8c23d67503821
• You are auditing DoneBench task quality for one domain. Workspace: D:\Researc… - codex-user-prompt:9d9df849dc02a87f
• app workf… - codex-user-prompt:39e6611a43c0e26f
• Read-only task in D:\Research\Prompt wiki. Inspect the current Prompt Wiki co… - codex-user-prompt:0c8fc0ba998bee5c
• Read-only user modeling task for Prompt Wiki. Infer realistic target users an… - codex-user-prompt:f7f5d729ae6f9c13
• 目前这个link graph的图颜值太低，且缺乏动态，需要重新修改 - codex-user-prompt:2515dd78d5e90a5f
• Repository: - codex-automation-prompt:daily-tgl-rec-research-review
• R2-real-LLM subgate provider is now approved. - codex-user-prompt:de0fdd7a5a32897b
• 当前状态： - codex-user-prompt:4f19015bf6dd7d8d
• Context from my IDE setup: - codex-user-prompt:4d075c616684094b
• 作为reviewer，对这个项目进行顶会级审稿，然后再多agent协同工作，把这个项目搞到顶会级别，必要时也可查阅其他顶会论文的setting，把你认为应… - codex-user-prompt:2c34f26f67a2c8d8
• 顶会setting explorer。这个项目名是 UncertaintyProtein-AI4S，目录 protein_bo_confor… - codex-user-prompt:3a5b1353e403cab1
• 工作区：D:\Research\DoneBench。你不是独自在代码库中工作，请不要回退他人的改动。当前项目是 DoneBench benchmark M… - codex-user-prompt:776fdd4df3f88dc7
• 工作区：D:\Research\DoneBench。你不是独自在代码库中工作，请不要回退他人的改动。当前项目是 DoneBench benchmark M… - codex-user-prompt:de8c7cf95961d2f8
• 最关键不是继续跑 API，而是去模板化任务生成，让 300 tasks 的内容多样性也真正对齐顶会。当前模块骨架已经像了，内容审计也诚实暴露了风险。现在还… - codex-user-prompt:de54a7595ceb8302
• 顶会相关工作差距分析。请联… - codex-user-prompt:e61e0cf823d613e0
• 工作区 D:\Research\DoneBench。你不是独自在代码库中工作，不要回退他人改动。任务：专注于去模板化任务生成质量。阅读 donebench… - codex-user-prompt:9e0ce10d574ddb71
• 看看我们还差哪些模块，建议直接全部补齐。 - codex-user-prompt:de2594e500941789
• Human Base… - codex-user-prompt:8d140d88dafd1b24
• 工作区 D:\Research\DoneBench。你不是独自在代码库中工作，不要回退他人改动。任务：补统计与错误 taxonomy 模块。职责范围：do… - codex-user-prompt:ab959833e00b8406
• Please audit D:\Research\DoneBench for benchmark-module completeness. Do not… - codex-user-prompt:903f4f1511e26676
• Please act as Reviewer 2 for D:\Research\DoneBench. Do not edit files. Compar… - codex-user-prompt:0296024d0a192f2d
• plans should… - codex-user-prompt:f38a91585edc14c5
• … - codex-user-prompt:f7c8b8dd66a4dd66
• external_project_baseline_packets.md 和 LLM4R… - codex-user-prompt:e5180275a57da774
• qwen3_lora_co… - codex-user-prompt:3c0fbdfba4c1e064
• tests for baseline provenance and… - codex-user-prompt:997d6c0f7262951c
• In D:\Research\TGL-Rec, inspect the current project docs and produce a detail… - codex-user-prompt:087ea905bb37bf9e
• *.md for outdated wording ar… - codex-user-prompt:67401ccd6cb7b8cc
• Audit D:\Research\TGL-Rec for stale wording after the new official baseline c… - codex-user-prompt:efa018abaaf54d50
• config locations. Fo… - codex-user-prompt:0734f3a1ad9d7aad
• 对于整个项目进行一次大改，保留几个重要的milestone，week1-4，pony12，light系列，shadow系列，baseline，从小域到四大… - codex-user-prompt:08be80405062fb9e
• You own the milestone architecture only. Read the current repo docs and draft… - codex-user-prompt:b9a3c8948d160985
• You own the top-conference review pass only. Use official conference sources… - codex-user-prompt:eac70cefcfd63480
• You own the reviewer persona pass only. Act like a top-conference reviewer re… - codex-user-prompt:099052ce642c17ef
• baselin… - codex-user-prompt:c1fbb4c9d766bfe1
• NeurIPS 风格）只读评审 D:\Research\TRUCE-Rec 当前研究工程路线。输出：可… - codex-user-prompt:d944fe05f732ac53
• week8_lar… - codex-user-prompt:42c36e9b904fced0
• Read-only task for TGL-Rec. Act as a strict top-conference reviewer. Inspect… - codex-user-prompt:bd46f6cf993cd3eb
• Read-only task. Act as a strict top-conference reviewer for TGL-Rec’s current… - codex-user-prompt:0408abebc4fb4c1a
• Read-only task. Audit D:\Research\TGL-Rec for server-run blockers in the full… - codex-user-prompt:e1bd08287b7c2746
• Related-work agent。请在 D:\Research\DoneBench 中只做只读分析，并结合公开顶会 agent… - codex-user-prompt:a21bc01f8b13cb33
• Execution agent。请在 D:\Research\DoneBench 中只读分析当前真实 tool-plan exec… - codex-user-prompt:d936c2c10c885dd4
• Artifact agent。请在 D:\Research\DoneBench 中只读分析 configs、CLI、repor… - codex-user-prompt:881474b76b729526
• Repository: D:\Research\TGL-Rec. Narrow review task while parent edits: inspe… - codex-user-prompt:4d41bfc3ffb05d1c
• 采用多agent协作方式，看看我们目前的framework设计是否足够强大，要原创不缝合，同时性能和复杂程度要上来很多，然后看看能否打败这些baselin… - codex-user-prompt:8f25e9708b8fc2ac
• 学长的建议：关于如何比较baseline是公平的，其实不同paper的定义不同，只要看起来你的标准是统一的就可以，我可以举几个例子 - codex-user-prompt:92df1b33c5d14a8e
• You own the fairness-policy design review. Based on the user’s senior advice,… - codex-user-prompt:d60194f3db48cef5
• config audit for implementing the new fairness policy. Inspe… - codex-user-prompt:2637b4df1ae9c5b0
• paper_claims_and_… - codex-user-prompt:a8f7b6c92a402045
• 实验设置审计员，快速检查当前仓库关于 external baseline fairness 的改动方向是否还有明显口径风险。重点看 RE… - codex-user-prompt:a304d778f43c0dd9
• provenance… - codex-user-prompt:d5b66dfbb1e0f372
• In repo D:\Research\TRUCE-Rec, inspect the Ours framework code and docs. Judg… - codex-user-prompt:23bdba74a880b055
• docs and four-domai… - codex-user-prompt:71d9da56d95831b7
• Repository: D:\Research\TGL-Rec. Audit what must change so server-side runs f… - codex-user-prompt:f27adfd965356006
• configs for baseline fairness po… - codex-user-prompt:b9ef71749c5072b3
• report，并自动生成一份 pilot_findings.md： - codex-user-prompt:ee032e74c8e9aba7
• paper structure and tel… - codex-user-prompt:23a96bc400f5d3ae
• In D:\Research\DoneBench, inspect audit-related modules and configs. Find the… - codex-user-prompt:264f195604df8c5d
• full-run readiness: configs, par… - codex-user-prompt:e523b5077ce737a6
• prepar… - codex-user-prompt:f87b64b58be9a705
• 在 D:\Research\Uncertainty 仓库中检查 main_make_official_external_adapter_plan.py、m… - codex-user-prompt:f0ad343d989437c9
• miles… - codex-user-prompt:71d5d8b67a2e5a87
• official runner 相关文档和脚本，重点找… - codex-user-prompt:2d03b3b845895c93
• gene… - codex-user-prompt:912a7b2434c5f986
• official_external_… - codex-user-prompt:b9d606d23d938f0a
• 报告属于 legac… - codex-user-prompt:2a5ab3701d78adb2
• server_runbook.md、OFFICIAL_EX… - codex-user-prompt:2db4cbc33f661e58
• Auditor agent. In this repo, evaluate whether the curren… - codex-user-prompt:52a7e0cce0ba1bd8
• c… - codex-user-prompt:30402d1703b51f98
• Protocol Scout. Use only local repository files unless… - codex-user-prompt:cb4aff10b848b272
• auditor 审核当前未提交的 SRPD formal 配置与… - codex-user-prompt:b73843ae79780526
• 就做ccrp和srpd这两个好了，我认为从观察（观察阶段基本就是beauty全域+其他三个域的各10000user（或者最好就是观察的数量和后面train… - codex-user-prompt:c366f0e1cf7418d4
• valid teacher 就不跑，不允许偷… - codex-user-prompt:4636540f704376cd
• 你是项目文档梳理 agent。只读梳理 D:\Research\DoneBench 的已有文档和代码结构，为后续 Codex handoff 文档提供事实… - codex-user-prompt:677d969b78e6bf83
• You are auditing DoneBench task quality for one domain. Workspace: D:\Researc… - codex-user-prompt:ba00b8d35ae19612
• You are auditing DoneBench task quality for one domain. Workspace: D:\Researc… - codex-user-prompt:fc87385eaaaf9512
• You are auditing DoneBench task quality for one domain. Workspace: D:\Researc… - codex-user-prompt:bc63fbd3dfc98e4d
• You are auditing DoneBench task quality for one domain. Workspace: D:\Researc… - codex-user-prompt:cdd80fca2d5b4419
• You are auditing DoneBench task quality for one domain. Workspace: D:\Researc… - codex-user-prompt:62189a1bfef60dc7
• You are auditing DoneBench task quality for one domain. Workspace: D:\Researc… - codex-user-prompt:529cdcd2c8a8d4bc
• You are auditing DoneBench task quality for one domain. Workspace: D:\Researc… - codex-user-prompt:41026be2af97c765
• You are auditing DoneBench task quality for one domain. Workspace: D:\Researc… - codex-user-prompt:1e30685ca6c788de
• AI benchmark 论文审稿人视角审这个项目：先读… - codex-user-prompt:5300686e01b28643
• agent_handoff.md 和 pa… - codex-user-prompt:a8f28f42af9c6262
• 最值得补的是 full corpus strict validation、oracle-spec ablation、token-matched ablat… - codex-user-prompt:dbd9268a3e42d1b4
• 报告侦察。目标：找出已有的 oracle-spec、token-matched、cross-f… - codex-user-prompt:a14158167ab588a4
• results.tex、experiments.t… - codex-user-prompt:1700179aaa56c40d
• PHASE_HANDOFF.md… - codex-user-prompt:4716cf0a2ce60e37
• project framework na… - codex-user-prompt:bf245358f3deb7e0
• PROJE… - codex-user-prompt:c0aa0dfebd31cf22
• ba… - codex-user-prompt:34252dec931467b9
• PROJECT_MEMOR… - codex-user-prompt:4b15f3cf1a4dbd76
• codex_project_memory.md, doc… - codex-user-prompt:0bfac25dcabeea2d
• phase… - codex-user-prompt:d83a69ed778bcd51
• *.tex、r… - codex-user-prompt:5c9de60b7c938f68
• *.tex 结构… - codex-user-prompt:b6e8fc6e92a569cd
• Kimi cross-family sli… - codex-user-prompt:d0f53be992730a1d
• 我希望能在agent这边加一句，每次碰到复杂任务都要用多agent协作工作，工作开始时默认阅读几个重要文档（这个要你自行决定和搜索你觉得重要的文档，一般是… - codex-user-prompt:98a479c6b788a20b
• agent 应该优先阅读哪些项目文档、长期操作规约应该放在哪些文件最有效。重… - codex-user-prompt:99c79f1078ae8eaf
• 在 D:\Research\DoneBench 中只读做顶会 reviewer 视角审查：用户想加一条规则，复杂任务必须多agent协作、开工默认读关键文… - codex-user-prompt:a83f386568cf5bc4
• … - codex-user-prompt:8e6885e290493b40
• SIGIR… - codex-user-prompt:1d5304b252568051
• Auditor，强势挑刺：我们自己的… - codex-user-prompt:8b860b1c625a0040
• RESEARCH… - codex-user-prompt:cec89807cc7d817a
• protocol scout for this repository. User asks whether… - codex-user-prompt:23abf0f842656324
• auditor for this repository. Review the c… - codex-user-prompt:9c7042b90c806b2a
• codex_project_memory.m… - codex-user-prompt:0d33f28e8468a1eb
• … - codex-user-prompt:3b819a8949c74bbe
• codex_project_memory.md and inspect local file… - codex-user-prompt:066dafd0c49ec61d
• 在 D:\Research\DoneBench 中只读梳理当前项目已有能力和缺口，目标是回答：除了 benchmark 贡献，还能自然长出哪些非缝合、非照… - codex-user-prompt:90cdd9fb8c024581
• area chair 视角：如果 DoneBench 不只作为 bench… - codex-user-prompt:7621309f64f2bbef
• advanced_stats… - codex-user-prompt:3d761dcf2aedb13d
• 你是 DoneBench 的 skeptical reviewer。请只读审查 M6.1 诊断表计划的 claim 风险：四象限表、自我违反 taxono… - codex-user-prompt:11c113201accfa0b
• agent_… - codex-user-prompt:c59967b9b0543b8a
• formal-verification 相关论文和项目：Age… - codex-user-prompt:76a27b2c15ef5701
• 报告状态，聚焦本轮计划的实现风险：M6.2 analysis.tex 写法、repaired diag… - codex-user-prompt:efb9e75078fcae34
• task-sampling reviewer for the DoneBenc… - codex-user-prompt:2e77b97419c0ee75
• analog-agent re… - codex-user-prompt:834558664795bf05
• Read-only task. Inspect D:\Research for Uncertainty-related research projects… - codex-user-prompt:0ac812da9b7e2c06
• Read-only task. Inspect D:\Research for truce-rec, tgl-rec, and donebench rel… - codex-user-prompt:2e82ff84bc08aa25
• For D:\Research\vipin’s knowledgebase, inspect AGENTS.md, WORKFLOWS.md, reade… - codex-user-prompt:85fd8fa1055d9513
• 文档审计，不要修改文件。目标：… - codex-user-prompt:c40559bb3a274a05
• 你是并行科研定位审计子 agent。请只读仓库 D:\Research\UncertaintyProtein-AI4S，不要修改文件。目标：从科研叙事角度… - codex-user-prompt:edc989d0ae27cc09
• 而且你的baseline目前也比较单薄，没有比较新的25年，26年出的那些顶会论文项目baseline，baseline一般至少要有8个，这些你不要口头说… - codex-user-prompt:0265e11e8c560a10
• 你是顶会 reviewer 子 agent，只读仓库 D:\Research\UncertaintyProtein-AI4S，不要修改。基于用户最新要求进… - codex-user-prompt:df18d1bedaa2b600
• 你是实现设计子 agent。请在你自己的工作区修改文件，注意你不是唯一 agent，不要撤销别人改动。任务：为 D:\Research\Uncertain… - codex-user-prompt:91f3ee293bf9f680
• observation也可最先从调用普通deepseek的api开始，然后再是上服务器用小模型去观察，发现确实有这个现象，然后我们就lora finetu… - codex-user-prompt:9c081859b7858b00
• deepseek要并行，qwen3 8B也要考虑调用并行，怎么快怎么来，但不影响实验，项目质量。 - codex-user-prompt:c7cc5a8352df7386
• 我想利用harness engineering 的全套思路（自己去搜寻里面包含什么概念），重新构建这个项目，内容idea只做部分保留，给你大的权限删去已有… - codex-user-prompt:c8feeb8a0b3abefd
• 我希望能在上服务器之前，本地先把所有已经能搭的项目都搭好，做好服务器上只需要跑指令即可。我还看到本地这个文件夹有很多东西是多余的可删的，你一定要大胆的做删… - codex-user-prompt:498e14cac586adae
• Plan-mode read-only audit only. Repo: D:\Research\UncertaintyProtein-AI4S. Us… - codex-user-prompt:32f15e8990e16635
• Plan-mode read-only research reviewer. Repo: D:\Research\UncertaintyProtein-A… - codex-user-prompt:dd7112df34ef630b
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:7ab57db4cc0ce8cc
• 本地再清一轮旧结构 - codex-user-prompt:a59c95f7a7254839
• smoke 旧配置、pap… - codex-user-prompt:69525abd1d8bd103
• few-shot optim… - codex-user-prompt:75d9a7d7c08332e4
• 做 server-ready adapter skeleton：为 registry 里优先级最高的 ProSpero、DPLM-2、GLID2E、ESM… - codex-user-prompt:5349e92c0fac7f1b
• 如果你还可用，请做只读 reviewer 审计：我们要把 ESM3 API 和 METL adapters 从 blocker skeleton 推进到真… - codex-user-prompt:ea5baa990066559d
• 2014… - codex-user-prompt:0d75cd686a79dbf4
• Repository: D:\Research\UncertaintyProtein-AI4S. Plan-mode only. Do not edit… - codex-user-prompt:04e3aaaa98a3fba2
• Repository: D:\Research\UncertaintyProtein-AI4S. Plan-mode only. Do not edit… - codex-user-prompt:e1eb09e28541f256
• Read-only audit task in D:\Research\UncertaintyProtein-AI4S. The user locked… - codex-user-prompt:1164d1ee329f43e7
• Read-only audit task in D:\Research\UncertaintyProtein-AI4S. Please inspect s… - codex-user-prompt:ace4e05c41f6eba6
• 在 D:\Research\Agent-AI4EDA\analog-agent 中快速检查 world model 相关代码结构，回答：1) 适合新增 d… - codex-user-prompt:61ad48a16badabc1
• 请作为 reviewer agent 对 D:\Research\Agent-AI4EDA\analog-agent 当前未提交 diff 做只读代码审查… - codex-user-prompt:d69057177bf964bb
• daily-arXiv-ai-… - codex-user-prompt:8983321fc3c7f7e5
• 论文发现产品的产品信息架构和可本地化功能：Papers with Code, Hugging Face T… - codex-user-prompt:5d2e0ba581239dd0
• 编码 agent”。仓库在 D:\Research\Research Recon Pro… - codex-user-prompt:c5d6bff537bb137e
• 你是 Research Recon Prototype 的“评审 agent”。仓库在 D:\Research\Research Recon Protot… - codex-user-prompt:1a3717f3f138aaee
• 知识库维护 agent”。仓库在 D:\Research\Research Recon… - codex-user-prompt:6896bb65287b0e91
• 清理 agent”。仓库在 D:\Research\Resea… - codex-user-prompt:548c8051c3db87fc
• 你是 Research Recon Prototype 的项目 reviewer。请在仓库 D:\Research\Research Recon Prot… - codex-user-prompt:eb9851b3d49f40ed
• 你是 Research Recon Prototype 的实现 worker，但先不要动手改文件。请在仓库 D:\Research\Research Re… - codex-user-prompt:403927342fccf060
• Code… - codex-user-prompt:f12042b38ec7d588
• 外部用户第一… - codex-user-prompt:95a712a86cff9434
• GitHub res… - codex-user-prompt:8130036803f10d39
• 你是 reference research agent B。请不要修改文件。请从竞品和成熟产品角度搜寻：Elicit、Semantic Scholar、C… - codex-user-prompt:c8f851e6a7105ea4
• PI，不只是看界面。工作目录是 D:\Research\Research Recon Proto… - codex-user-prompt:fac1ae12f847bcb8
• Codex automation user，要像用户一样实际跑端到端，不只是… - codex-user-prompt:20b97fb7478d5966
• 你是 QA Agent 6。先不要改代码。工作目录 D:\Research\Research Recon Prototype。任务： - codex-user-prompt:5bf42e552b131cd9
• 追加 reviewer 发现，请纳入你的后端实作范围： - codex-user-prompt:1ca6f85b01aa68b2
• A… - codex-user-prompt:223ebc8aba98d87d
• 验证 agent”，检查现有测试与 eval… - codex-user-prompt:8ea469fe5bd89574
• 只读规划任务。仓库路径 D:\Research\Research Recon Prototype。作为“评审 agent”，审查给 Research Re… - codex-user-prompt:e0ee370592152c00
• 知识库维护 agent”，检查 README… - codex-user-prompt:37d804e8c7353000
• 清理 agent”… - codex-user-prompt:0c9a527257694c53
• 只读规划任务。仓库路径 D:\Research\Research Recon Prototype。作为“故障修复 agent”，预判实现 LLM4Rec… - codex-user-prompt:05c40f59c4ff1809
• fairness reviewer pass for a pro… - codex-user-prompt:50efe2c7d46f746c
• Use $truce-rec at D:\Research\TRUCE-Rec.codex\skills\truce-rec to prepare fo… - codex-user-prompt:6bd47058f9460e1b
• In D:\Research\Uncertainty, inspect main_train_score_setrec_upstream_adapter… - codex-user-prompt:ea33c6b4a9e5a006
• auditor pass for the SETRec o… - codex-user-prompt:1a34c4388d2e6051
• In D:\Research\Uncertainty, inspect the current SETRec official adapter code,… - codex-user-prompt:aac1d1a20b9f26f4
• In D:\Research\Uncertainty, focus on memory mitigation for the SETRec officia… - codex-user-prompt:99f2eccf5d746691
• training code… - codex-user-prompt:09f54fd46dc0b387
• auditor for the pasted SETRec off… - codex-user-prompt:6131c215c59fc34f
• 我感觉这个baseline错误好像特别多，我们也改了很久了，没啥效果。这样吧，我们做其他的学长推荐的baseline吧，然后另外我不知道你记不记得之前你找… - codex-user-prompt:964c7257489e914a
• [[sources/codex-prompts/codex-user-prompt-f95d989c872c4ba4|elmrec-official] epoch=75 train_loss=0.022439]] - codex-user-prompt:f95d989c872c4ba4
• [[sources/codex-prompts/codex-user-prompt-75ead59f47dfc24b|[elmrec-official] epoch=75 train_loss=0.022439]] - codex-user-prompt:75ead59f47dfc24b
• [[sources/codex-prompts/codex-user-prompt-453e81fdc20c4306|[proex-official] epoch=2970 train_loss=7.186873]] - codex-user-prompt:453e81fdc20c4306
• ELMRec official runner 的命令面、输出… - codex-user-prompt:e2066354f40dc8a4
• auditor 只读检查：针对用户当前服务器上 book 域第七个 off… - codex-user-prompt:cc408adbf0dd9ed9
• [[sources/codex-prompts/codex-user-prompt-6d3b1656dd0d215e|[promax-official] epoch=2980 train_loss=1.700973]] - codex-user-prompt:6d3b1656dd0d215e
• 跑full的话如果说他能做到中途中断还能接续，就跑。如果不能，就先做到这点，然后再跑。还有跑的时候要间断性的报告跑了多少，还要跑多久等，但不必过分频繁。高… - codex-user-prompt:bf803d25f570bcd6
• 跑full的话如果说他能做到中途中断还能接续，就跑。如果不能，就先做到这点，然后再跑。还有跑的时候要间断性的报告跑了多少，还要跑多久等，但不必过分频繁。高… - codex-user-prompt:6ca099be41f4c475
• repro checker。请检查如何实现 claim_to_artifact_map、leaderboard_contamina… - codex-user-prompt:d6fd98533dcff877
• Plan-mode exploration only: do not edit files. Workspace: D:\Research\vipin’s… - codex-user-prompt:8aaa918c83611f31
• In workspace D:\Research\vipin’s knowledgebase, do read-only exploration for… - codex-user-prompt:ed25a7d3da4e0e76
• d盘下的academic-portfolio,undergraducate-project-netherlands,undergraduate-study… - codex-user-prompt:a3146fab9498e0f5
• Read-only exploration. Inspect D:\academic-portfolio and D:\weipingyan-portfo… - codex-user-prompt:85ab75319b6d44c7
• Read-only exploration. Inspect D:\undergraducate-project-netherlands and D:\u… - codex-user-prompt:4633bf483676282f
• knowledge-gr… - codex-user-prompt:9fed3741f5dc3052
• In D:\Research\vipin’s knowledgebase, inspect the local vipin wiki site publi… - codex-user-prompt:f502f65fa0227372
• Files mentioned by the user: - codex-user-prompt:38de5993c7dc6194
• Files mentioned by the user: - codex-user-prompt:25511c6845f07781
• Agent 2: Read-only inspect Prompt Wiki prompt assets and docs. Focus on conte… - codex-user-prompt:8c218b00fbabd7bb
• stati… - codex-user-prompt:89c9f5b5e4b5d61f
• Agent 5: Apply OpenAI harness-engineering mindset conceptually to Prompt Wiki… - codex-user-prompt:f12610877d683317
• 我建议仿照vipin-wiki来做，特别是前端，但别动那个项目的代码，大胆删去我们这边多余的一些文件。但我们的内核仍然是prompt-wiki，所以原材料… - codex-user-prompt:86db633112406c1e
• Read-only exploration task for Prompt Wiki. Do not modify files. In D:\Resear… - codex-user-prompt:d2830e19aca38bcd
• Read-only exploration task for Prompt Wiki. Do not modify files. In D:\Resear… - codex-user-prompt:2b67a78f4a8b4661
• Read-only exploration task for vipin reference project. Do not modify files… - codex-user-prompt:5b8934d4af7fd78f
• Read-only exploration task. Compare Prompt Wiki content assets in D:\Research… - codex-user-prompt:455136be71870ba1
• Read-only exploration task for cleanup planning. Do not modify files. In D:\R… - codex-user-prompt:f36fdafc41743906
• prompts, packages… - codex-user-prompt:987e43ff1c700b41
• public sources if available, benchmark top adja… - codex-user-prompt:0f9a76c50c6deec9
• Read the docs folder first. - codex-user-prompt:72cebbbe4b664fef
• 更新到wiki - codex-user-prompt:e06bc73154199aa8
• [[sources/codex-prompts/codex-user-prompt-46875507a5030d77|https://developers.openai.com/cookbook 这个你要仔细ingest，这个对我过分重要，主要就是得分好类，然后具体说说讲…]] - codex-user-prompt:46875507a5030d77
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:e4f43c728942913c
• 立党老师他的所有公开发布信息，主要是X上和youtube上，这个你要仔细ingest，这个对我过分重要，主要就是得分好类，然后具体说说讲了什么。然后之后这… - codex-user-prompt:be10f475d897d016
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:6d46b284910d5ef4
• X 官方 profile 可访问，但未直接暴露 status 链接；3 个辅助 mirror 当前不可达，已作为 3 条 crawl errors 记录进… - codex-user-prompt:f588ceb0f4571c69
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:8e8c4e6f90663c30
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:193eb2f7cef5392e
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:797e551aa4577a62
• 你应该能看到我skill文件夹下的一些文件吧，这些都是从别人的仓库移植进来的，但你现在要做的事就是追溯原始仓库，这个归类应该为skill，然后ingest… - codex-user-prompt:2748a5a90b2cac88
• 你应该能看到我本地codex的所有我对codex的prompt吧，这个对我过分重要，我希望能让wiki ingest，分好类，定时自动爬取加更新东西。然后… - codex-user-prompt:840448f9bd1b24f6
• PLEASE IMPLEMENT THIS PLAN: - codex-user-prompt:bed42e1493ef08d3
• Run the weekly update for the selected local Codex prompt corpus in the vipin… - codex-automation-prompt:update-codex-prompt-corpus
• Run the weekly research inspiration digest in the vipin wiki repository. From… - codex-automation-prompt:weekly-research-inspiration-digest
• Run the weekly digest for the Lidang public ideas corpus in the vipin wiki re… - codex-automation-prompt:digest-lidang-public-ideas
• Run the weekly update for the Shunyu Yao public corpora in the vipin wiki rep… - codex-automation-prompt:update-shunyu-yao-public-corpora
• Run the weekly update for the frontend project frameworks public corpus in th… - codex-automation-prompt:update-frontend-frameworks-public-corpus
• Run the Karpathy public corpus ingest workflow in the vipin wiki repository… - codex-automation-prompt:update-karpathy-public-corpus
• vipin’s… - codex-automation-prompt:weekly-openai-cookbook-wiki-ingest

Recently Crystallized

• 2026-06-19-weiping-wiki-upgrade-audit — WEIPING_WIKI Upgrade Audit (2026-06-19)
• 2026-06-19-llm-wiki-v2-pattern — LLM Wiki v2 source note (2026-06-19)