# Send Virtual Reading Group to your agent Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually. ## Fast path - Download the package from Yavira. - Extract it into a folder your agent can access. - Paste one of the prompts below and point your agent at the extracted folder. ## Suggested prompts ### New install ```text I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete. ``` ### Upgrade existing ```text I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run. ``` ## Machine-readable fields ```json { "schemaVersion": "1.0", "item": { "slug": "virtual-reading-group", "name": "Virtual Reading Group", "source": "tencent", "type": "skill", "category": "AI 智能", "sourceUrl": "https://clawhub.ai/IsonaEi/virtual-reading-group", "canonicalUrl": "https://clawhub.ai/IsonaEi/virtual-reading-group", "targetPlatform": "OpenClaw" }, "install": { "downloadUrl": "/downloads/virtual-reading-group", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=virtual-reading-group", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "packageFormat": "ZIP package", "primaryDoc": "SKILL.md", "includedAssets": [ "SKILL.md", "assets/synthesis-template.md", "references/default-personas.md", "references/paper-notes-template.md", "references/workflow.md" ], "downloadMode": "redirect", "sourceHealth": { "source": "tencent", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-04-30T16:55:25.780Z", "expiresAt": "2026-05-07T16:55:25.780Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentDisposition": "attachment; filename=\"network-1.0.0.zip\"", "redirectLocation": null, "bodySnippet": null }, "scope": "source", "summary": "Source download looks usable.", "detail": "Yavira can redirect you to the upstream package for this source.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/virtual-reading-group" }, "validation": { "installChecklist": [ "Use the Yavira download entry.", "Review SKILL.md after the package is downloaded.", "Confirm the extracted package contains the expected setup assets." ], "postInstallChecks": [ "Confirm the extracted package includes the expected docs or setup files.", "Validate the skill or prompts are available in your target agent workspace.", "Capture any manual follow-up steps the agent could not complete." ] } }, "links": { "detailUrl": "https://openagent3.xyz/skills/virtual-reading-group", "downloadUrl": "https://openagent3.xyz/downloads/virtual-reading-group", "agentUrl": "https://openagent3.xyz/skills/virtual-reading-group/agent", "manifestUrl": "https://openagent3.xyz/skills/virtual-reading-group/agent.json", "briefUrl": "https://openagent3.xyz/skills/virtual-reading-group/agent.md" } } ``` ## Documentation ### Virtual Reading Group Orchestrate parallel expert agents to read papers, discuss findings, challenge each other's interpretations, and synthesize an integrated discussion document with traceable citations. ### Quick Start Minimum inputs required: Research question — the lens through which papers are analyzed Paper list — paths to PDFs/text files, or paper descriptions for web lookup Output directory — where all outputs are written Optional inputs: Custom expert personas (default: see references/default-personas.md) Custom junior researcher persona Language preference (default: English) Number of experts (default: auto-calculated from paper count) ### Workflow Overview The skill runs 4 sequential phases. Each phase must complete before the next begins. PhaseAgentsInputOutput1. Paper ReadingN experts (parallel)Papers + research question{AuthorYear}_notes.md, {Expert}_session_summary.md2. Junior Discussion1 junior researcherAll Phase 1 outputs{Junior}_discussion.md3. Expert ResponsesN experts (parallel)Phase 2 output + other experts' summaries{Expert}_response_to_{Junior}.md4. Synthesis1 synthesizerAll previous outputsIntegrated_Discussion_Summary.md For detailed prompts and phase specifications: Read references/workflow.md. ### Orchestration Procedure ⚠️ Important: The prompts below are abbreviated summaries. For full prompt templates that produce quality output, use references/workflow.md. The pseudocode blocks show orchestration structure — adapt to your actual sub-agent spawning mechanism. ### 1. Validate Inputs - Confirm research question is specified - Confirm paper list is non-empty - Confirm output directory exists or create it - Load personas from user input or references/default-personas.md ### 2. Calculate Expert Assignment Determine number of experts and paper batches: if paper_count <= 4: num_experts = 1 elif paper_count <= 10: num_experts = 2 elif paper_count <= 20: num_experts = min(4, ceil(paper_count / 5)) else: num_experts = min(8, ceil(paper_count / 5)) Distribute papers evenly across experts (max 5 per expert). # ⚠️ Context contamination warning: assigning >5 papers per expert degrades # note quality — later papers in the batch get shallower treatment as context # fills up. Prefer 3-5 papers per agent for best results. ### 3. Execute Phase 1 — Paper Reading (Parallel) For each expert, spawn a sub-agent with: Label: expert-reader-{expert_name} Model: opus (or sonnet for budget) Core instructions: Read assigned papers through research question lens Write notes using references/paper-notes-template.md Save as {output_dir}/{AuthorYear}_notes.md Write session summary with cross-cutting themes Critical: Quote specific passages with section labels — all claims must be traceable 📄 Full prompt template: See references/workflow.md → Phase 1 Wait for all Phase 1 agents to complete before proceeding. ### 4. Execute Phase 2 — Junior Discussion (Single Agent) Spawn single agent with: Label: junior-discussion Model: opus (required — needs strong reasoning) Core instructions: Read all Phase 1 outputs (notes + summaries) For each paper: summarize claims, pose challenging questions to each expert Generate Grand Questions: 3 unsolved problems, 2 testable hypotheses, 2 methodological gaps Reference specific passages — be intellectually provocative 📄 Full prompt template: See references/workflow.md → Phase 2 Wait for Phase 2 to complete before proceeding. ### 5. Execute Phase 3 — Expert Responses (Parallel) For each expert, spawn a sub-agent with: Label: expert-response-{expert_name} Model: opus (recommended) Core instructions: Read junior's discussion + other experts' summaries + own notes Respond to each question directed at them (150-300 words per response) Reference specific paper passages, engage with other expert's perspective Respond to Grand Questions from their domain expertise Be collegial but intellectually rigorous — disagree where warranted 📄 Full prompt template: See references/workflow.md → Phase 3 Wait for all Phase 3 agents to complete before proceeding. ### 6. Execute Phase 4 — Synthesis (Single Agent) Spawn single agent with: Label: synthesis Model: opus (required — complex reasoning) Core instructions: Read ALL files from Phases 1-3 Follow assets/synthesis-template.md structure Organize by THEME, not by paper or speaker Every claim attributed: [Expert_A]/[Expert_B]/[Junior] + (PaperCode, §Section) Include: Points of Consensus, Points of Disagreement, Open Questions Synthesize, don't summarize — find the intellectual threads 📄 Full prompt template: See references/workflow.md → Phase 4 ### 7. Report Completion List all generated files and provide a brief summary of the discussion themes. ### Deeper Discussion If user wants experts to expand on specific points: Spawn new expert response agent(s) with targeted follow-up questions Re-run Phase 4 synthesis including the additional responses ### Second Round For a full second round (new questions, new responses): Rename Phase 2-4 outputs with round suffix (e.g., Chen_discussion_r1.md) Re-run Phase 2 with instruction to build on previous round Continue through Phases 3-4 ### Recovery from Partial Run If a phase fails: Check error handling in references/workflow.md Retry failed agent(s) individually Continue from last successful phase (outputs are saved incrementally) ### File Naming Conventions File TypePatternExamplePaper notes{FirstAuthorLastName}{Year}_notes.mdChen2024_notes.mdExpert summary{ExpertLastName}_session_summary.mdLin_session_summary.mdJunior discussion{JuniorLastName}_discussion.mdChen_discussion.mdExpert response{ExpertLastName}_response_to_{JuniorLastName}.mdLin_response_to_Chen.mdSynthesisIntegrated_Discussion_Summary.md— ### Citation Requirements Enforce in all agent prompts: Every factual claim must reference a paper Use format: (AuthorYear, §Section) or (AuthorYear, p.X) Direct quotes must include section/page Discussion claims must attribute speaker: [Expert_A], [Expert_B], [Junior] ### ⚠️ Anti-Fabrication Rule (Critical) Never fabricate citations. If an agent cannot find the exact passage in the source text: Leave the field blank or write Do NOT paraphrase and present it as a quote Do NOT infer what the paper "probably says" Fabricated citations are worse than missing citations — they corrupt the knowledge base silently. Accuracy > Coverage. ### No Source = No Notes If a paper has no PDF or markdown source available: Write a placeholder note with status 📭 未讀 Leave all content sections blank Do NOT attempt to write notes from memory or web search results Only write substantive notes when the actual source document is accessible. ### Scaling Guidelines PapersExpertsBatchesEstimated Time1-61115-20 min7-122220-30 min13-243-43-430-45 min25-504-85-845-90 min ### Custom Personas Replace default personas by providing: Expert A: Dr. [Name], [Role]. Background in [X]. Emphasizes [methodology/perspective]. Skeptical of [Y]. Tone: [collegial/rigorous/provocative]. Expert B: Dr. [Name], [Role]. Background in [X]. ... See references/default-personas.md for complete templates. ### Language Pass the language parameter when invoking the orchestration: All agent prompts include Language: {language} instruction Agents read papers and write outputs in the specified language Default: English Example: "Run the reading group in Japanese" → adds Language: Japanese to all phase prompts. ### Model Selection Model choice significantly impacts output quality and cost: ConfigurationPhasesQualityCostUse WhenFull opusAll phases use opusHighest$$$Publication-quality analysis, complex papersMixedPhase 1: sonnet, Phases 2-4: opusHigh$$Good balance — reading is less reasoning-intensiveBudgetAll phases use sonnetMedium$Quick exploration, simpler papers Recommendations: Phase 2 (Junior Discussion) benefits most from opus — requires synthesizing multiple papers and generating non-obvious questions Phase 4 (Synthesis) also benefits from opus — thematic organization requires complex reasoning Phase 1 (Reading) can use sonnet if papers aren't highly technical Phase 3 (Responses) can use sonnet if questions are straightforward ### Integration This skill is standalone but works well with paper collection workflows: literature-manager or similar skills: Use to gather and organize papers first, then pass the collection to virtual-reading-group PDF extraction tools: Pre-extract text from PDFs if agents have trouble reading them directly ### References references/workflow.md — Detailed phase specifications and full prompt templates references/default-personas.md — Ready-to-use expert and junior researcher personas references/paper-notes-template.md — Template for individual paper notes ### Assets assets/synthesis-template.md — Structure for the final integrated discussion summary ## Trust - Source: tencent - Verification: Indexed source record - Publisher: IsonaEi - Version: 1.1.0 ## Source health - Status: healthy - Source download looks usable. - Yavira can redirect you to the upstream package for this source. - Health scope: source - Reason: direct_download_ok - Checked at: 2026-04-30T16:55:25.780Z - Expires at: 2026-05-07T16:55:25.780Z - Recommended action: Download for OpenClaw ## Links - [Detail page](https://openagent3.xyz/skills/virtual-reading-group) - [Send to Agent page](https://openagent3.xyz/skills/virtual-reading-group/agent) - [JSON manifest](https://openagent3.xyz/skills/virtual-reading-group/agent.json) - [Markdown brief](https://openagent3.xyz/skills/virtual-reading-group/agent.md) - [Download page](https://openagent3.xyz/downloads/virtual-reading-group)