{"id":"d958ae6e-fe7e-4018-adc0-7c65a4af56b8","entityType":"agent","slug":"clawhub-eric-promax-academic-pipeline","name":"academic-pipeline-v1","canonicalUrl":"https://www.xpersona.co/agent/clawhub-eric-promax-academic-pipeline","canonicalPath":"/agent/clawhub-eric-promax-academic-pipeline","generatedAt":"2026-10-11T20:59:20.177Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"editorial-content","verified":true,"confidence":"high","updatedAt":"2026-10-11T15:53:46.384Z","emptyReason":null},"description":"Orchestrator for the full academic research pipeline: literature search -> research -> write -> integrity check -> review -> revise -> re-review -> re-revise... Skill: academic-pipeline-v1 Owner: eric-promax Summary: Orchestrator for the full academic research pipeline: literature search -> research -> write -> integrity check -> review -> revise -> re-review -> re-revise... Tags: latest:1.0.0 Version history: v1.0.0 | 2026-05-14T02:46:10.557Z | user Academic Pipeline v1.0.0 — Initial Release - Introduces a full academic research workflow orchestrator covering 12 stages from","descriptionLabel":"Technical summary","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 1K downloads reported by the source. Last updated 10/11/2026.","installCommand":"clawhub skill install s1746e7bcag3t464dy9vye9kr586pe2v:academic-pipeline","sourceUrl":"https://clawhub.ai/eric-promax/academic-pipeline","homepage":"https://clawhub.ai/eric-promax/skills/academic-pipeline","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/eric-promax/academic-pipeline","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/eric-promax/skills/academic-pipeline","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":60,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"Orchestrator for the full academic research pipeline: literature search -> research -> write -> integrity check -> review -> revise -> re-review -> re-revise..."},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-11T15:53:46.384Z","emptyReason":null},"protocols":[{"protocol":"OPENCLEW","label":"OpenClaw","status":"self-declared","notes":"Declared in the public agent profile."}],"capabilities":[],"verifiedCount":0,"selfDeclaredCount":1,"capabilityMatrix":{"rows":[{"key":"OPENCLEW","type":"protocol","support":"unknown","confidenceSource":"profile","notes":"Listed on profile"}],"flattenedTokens":"protocol:OPENCLEW|unknown|profile"}},"adoption":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T15:53:46.384Z","emptyReason":null},"stars":null,"forks":null,"downloads":1034,"packageName":null,"latestVersion":"1.0.0","tractionLabel":"1K downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T15:53:46.301Z","emptyReason":null},"lastUpdatedAt":"2026-10-11T15:53:46.384Z","lastCrawledAt":"2026-10-11T15:53:46.301Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-12T15:53:46.301Z","lastVerifiedAt":null,"highlights":[{"version":"1.0.0","createdAt":"2026-05-14T02:46:10.557Z","changelog":"Academic Pipeline v1.0.0 — Initial Release - Introduces a full academic research workflow orchestrator covering 12 stages from literature search to final manuscript. - Coordinates multiple academic research and writing skills, including search, drafting, review, revision, and humanization. - Implements mandatory user confirmation checkpoints at every stage. - Integrates two-stage peer review and integrity verification for references and data. - Adds automated process documentation of the entire creation workflow. - Supports adaptive triggers, mid-entry, and parallelization opportunities for increased efficiency.","fileCount":25,"zipByteSize":96812}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s1746e7bcag3t464dy9vye9kr586pe2v:academic-pipeline","setupComplexity":"low","setupSteps":["Setup complexity is LOW. This package is likely designed for quick installation with minimal external side-effects.","Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data."],"contract":{"contractStatus":"missing","authModes":[],"requires":[],"forbidden":[],"supportsMcp":false,"supportsA2a":false,"supportsStreaming":false,"inputSchemaRef":null,"outputSchemaRef":null,"dataRegion":null,"contractUpdatedAt":null,"sourceUpdatedAt":null,"freshnessSeconds":null},"invocationGuide":{"preferredApi":{"snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-eric-promax-academic-pipeline/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-eric-promax-academic-pipeline/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-eric-promax-academic-pipeline/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-eric-promax-academic-pipeline/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-eric-promax-academic-pipeline/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-eric-promax-academic-pipeline/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-11T20:59:20.173Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-eric-promax-academic-pipeline/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-eric-promax-academic-pipeline/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-eric-promax-academic-pipeline/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-eric-promax-academic-pipeline/trust"}},"reliability":{"evidence":{"source":"runtime-metrics","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No trust, reliability, or runtime telemetry is available."},"trust":{"status":"unavailable","handshakeStatus":"UNKNOWN","verificationFreshnessHours":null,"reputationScore":null,"p95LatencyMs":null,"successRate30d":null,"fallbackRate":null,"attempts30d":null,"trustUpdatedAt":null,"trustConfidence":"unknown","sourceUpdatedAt":null,"freshnessSeconds":null},"decisionGuardrails":{"doNotUseIf":["Contract metadata is missing or unavailable for deterministic execution."],"safeUseWhen":[],"riskFlags":["missing_or_unavailable_contract","trust_data_unavailable","schema_references_missing"],"operationalConfidence":"low"},"executionMetrics":{"observedLatencyMsP50":null,"observedLatencyMsP95":null,"estimatedCostUsd":null,"uptime30d":null,"rateLimitRpm":null,"rateLimitBurst":null,"lastVerifiedAt":null,"verificationSource":null},"runtimeMetrics":{"successRate":null,"avgLatencyMs":null,"avgCostUsd":null,"hallucinationRate":null,"retryRate":null,"disputeRate":null,"p50Latency":null,"p95Latency":null,"lastUpdated":null}},"benchmarks":{"evidence":{"source":"no-benchmark-data","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No benchmark suites or observed failure patterns are available."},"suites":[],"failurePatterns":[]},"artifacts":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"high","updatedAt":"2026-10-11T15:53:46.384Z","emptyReason":null},"readme":"Skill: academic-pipeline-v1\n\nOwner: eric-promax\n\nSummary: Orchestrator for the full academic research pipeline: literature search -> research -> write -> integrity check -> review -> revise -> re-review -> re-revise...\n\nTags: latest:1.0.0\n\nVersion history:\n\nv1.0.0 | 2026-05-14T02:46:10.557Z | user\n\nAcademic Pipeline v1.0.0 — Initial Release\n\n- Introduces a full academic research workflow orchestrator covering 12 stages from literature search to final manuscript.\n- Coordinates multiple academic research and writing skills, including search, drafting, review, revision, and humanization.\n- Implements mandatory user confirmation checkpoints at every stage.\n- Integrates two-stage peer review and integrity verification for references and data.\n- Adds automated process documentation of the entire creation workflow.\n- Supports adaptive triggers, mid-entry, and parallelization opportunities for increased efficiency.\n\nArchive index:\n\nArchive v1.0.0: 25 files, 96812 bytes\n\nFiles: agents/integrity_verification_agent.md (24608b), agents/pipeline_orchestrator_agent.md (22039b), agents/state_tracker_agent.md (15854b), examples/full_pipeline_example.md (17176b), examples/integrity_failure_recovery.md (27029b), examples/mid_entry_example.md (12640b), references/ai_research_failure_modes.md (14978b), references/changelog.md (4984b), references/claim_verification_protocol.md (2784b), references/external_review_protocol.md (6329b), references/integrity_review_protocol.md (2390b), references/mode_advisor.md (8025b), references/pipeline_state_machine.md (13388b), references/plagiarism_detection_protocol.md (13795b), references/process_summary_protocol.md (11245b), references/progress_dashboard_template.md (1380b), references/reinforcement_content.md (1132b), references/reproducibility_audit.md (1642b), references/score_trajectory_protocol.md (3758b), references/team_collaboration_protocol.md (9127b), references/two_stage_review_protocol.md (1404b), skill-card.md (3304b), SKILL.md (27908b), templates/pipeline_status_template.md (3813b), _meta.json (136b)\n\nFile v1.0.0:SKILL.md\n\n---\nname: academic-pipeline\ndescription: \"Orchestrator for the full academic research pipeline: literature search -> research -> write -> integrity check -> review -> revise -> re-review -> re-revise -> final integrity check -> humanize -> finalize. Coordinates academic-search, deep-research, academic-paper, academic-paper-reviewer, and humanizer into a seamless 12-stage workflow with mandatory integrity verification, two-stage peer review, de-AI processing, and reproducible quality gates. Triggers on: academic pipeline, research to paper, full paper workflow, paper pipeline, end-to-end paper, research-to-publication, complete paper workflow.\"\nmetadata:\n  version: \"3.5\"\n  last_updated: \"2026-05-13\"\n  depends_on: \"ima-skills, academic-search, deep-research, academic-paper, academic-paper-reviewer, humanizer, humanizer-zh\"\n  status: active\n  related_skills:\n    - ima-skills\n    - academic-search\n    - deep-research\n    - academic-paper\n    - academic-paper-reviewer\n    - humanizer\n    - humanizer-zh\n---\n\n# Academic Pipeline v3.5 — Full Academic Research Workflow Orchestrator\n\nA lightweight orchestrator that manages the complete academic pipeline from research exploration to final manuscript. It does not perform substantive work — it only detects stages, recommends modes, dispatches skills, manages transitions, and tracks state.\n\n**v2.0 Core Improvements**:\n1. **Mandatory user confirmation checkpoints** — Each stage completion requires user confirmation before proceeding to the next step\n2. **Academic integrity verification** — After paper completion and before review submission, 100% reference and data verification must pass\n3. **Two-stage review** — First full review + post-revision focused verification review\n4. **Final integrity check** — After revision completion, re-verify all citations and data are 100% correct\n5. **Reproducible** — Standardized workflow producing consistent quality assurance each time\n6. **Process documentation** — After pipeline completion, automatically generates a \"Paper Creation Process Record\" PDF documenting the human-AI collaboration history\n\n## Quick Start\n\n**Full workflow (from scratch):**\n```\nI want to write a research paper on the impact of AI on higher education quality assurance\n```\n--> academic-pipeline launches, starting from Stage 2 (RESEARCH)\n\n**Mid-entry (existing paper):**\n```\nI already have a paper, help me review it\n```\n--> academic-pipeline detects mid-entry, starting from Stage 4 (INTEGRITY)\n\n**Revision mode (received reviewer feedback):**\n```\nI received reviewer comments, help me revise\n```\n--> academic-pipeline detects, starting from Stage 7 (REVISE)\n\n**Execution flow:**\n1. Detect the user's current stage and available materials\n2. Recommend the optimal mode for each stage\n3. Dispatch the corresponding skill for each stage\n4. **After each stage completion, proactively prompt and wait for user confirmation**\n5. Track progress throughout; Pipeline Status Dashboard available at any time\n\n---\n\n## Trigger Conditions\n\n### Trigger Keywords\n\n**English**: academic pipeline, research to paper, full paper workflow, paper pipeline, end-to-end paper, research-to-publication, complete paper workflow\n\n### Non-Trigger Scenarios\n\n| Scenario | Skill to Use |\n|----------|-------------|\n| Only need to search materials or do a literature review | `deep-research` |\n| Only need to write a paper (no research phase needed) | `academic-paper` |\n| Only need to review a paper | `academic-paper-reviewer` |\n| Only need to check citation format | `academic-paper` (citation-check mode) |\n| Only need to convert paper format | `academic-paper` (format-convert mode) |\n\n### Trigger Exclusions\n\n- If the user only needs a single function (just search materials, just check citations), no pipeline is needed — directly trigger the corresponding skill\n- If the user is already using a specific mode of a skill, do not force them into the pipeline\n- The pipeline is optional, not mandatory\n\n---\n\n## Pipeline Stages (12 Stages)\n\n| Stage | Name | Skill / Agent Called | Available Modes | Deliverables |\n|-------|------|---------------------|----------------|-------------|\n| **1** | **LITERATURE SEARCH** | **`academic-search`** | **multi-source, single-source, two-pass, query-expansion** | **Bibliography (BibTeX) + metadata (JSON) + CCF/venue level + optional PDFs** |\n| 2 | RESEARCH | `deep-research` | socratic, full, quick | RQ Brief, Methodology, Bibliography, Synthesis |\n| 3 | WRITE | `academic-paper` | plan, full | Paper Draft |\n| **4** | **INTEGRITY** | **`integrity_verification_agent`** | **pre-review** | **Integrity verification report + corrected paper** |\n| 5 | REVIEW | `academic-paper-reviewer` | full (incl. Devil's Advocate) | 5 review reports + Editorial Decision + Revision Roadmap |\n| **6** | **RE-REVIEW** | **`academic-paper-reviewer`** | **re-review** | **Verification review report: revision response checklist + residual issues** |\n| 7 | REVISE | `academic-paper` | revision | Revised Draft, Response to Reviewers |\n| **8** | **RE-REVISE** | **`academic-paper`** | **revision** | **Second revised draft (if needed)** |\n| **9** | **FINAL INTEGRITY** | **`integrity_verification_agent`** | **final-check** | **Final verification report (must achieve 100% pass to proceed)** |\n| **10** | **HUMANIZE** | **`humanizer` + `humanizer-zh`** | **full (双语互补)** | **De-AI 化论文全文 + 双语变更摘要** |\n| 11 | FINALIZE | `academic-paper` | format-convert | Final Paper (default MD + DOCX; ask about LaTeX; confirm correctness; PDF) |\n| **12** | **PROCESS SUMMARY** | **orchestrator** | **auto** | **Paper creation process record MD + LaTeX to PDF (bilingual)** |\n\n**Parallelization opportunity (v3.3)**: Within Stage 3, the `academic-paper` skill's Phase 1 (literature_strategist_agent) and the `visualization_agent` can operate in parallel after Phase 2 (structure_architect_agent) completes the outline. Specifically:\n- Once the outline includes a visualization plan, `visualization_agent` can begin figure generation\n- Simultaneously, `argument_builder_agent` can build CER chains\n- `draft_writer_agent` waits for both to complete before beginning Phase 4\n\nThis mirrors PaperOrchestra's parallel execution of Plot Generation (Step 2) and Literature Review (Step 3) after Outline (Step 1), which reduces overall pipeline latency. The parallelization is optional — sequential execution remains the default for simplicity.\n\n---\n\n## Pipeline State Machine\n\n1. **Stage 1 LITERATURE SEARCH** -> user confirmation -> Stage 2\n2. **Stage 2 RESEARCH** -> user confirmation -> Stage 3\n3. **Stage 3 WRITE** -> user confirmation -> Stage 4\n4. **Stage 4 INTEGRITY** -> PASS -> Stage 5 (FAIL -> fix and re-verify, max 3 rounds)\n5. **Stage 5 REVIEW** -> Accept -> Stage 9 / Minor|Major -> Stage 7 / Reject -> Stage 3 or end\n6. **Stage 7 REVISE** -> user confirmation -> Stage 6\n7. **Stage 6 RE-REVIEW** -> Accept|Minor -> Stage 9 / Major -> Stage 8\n8. **Stage 8 RE-REVISE** -> user confirmation -> Stage 9 (no return to review)\n9. **Stage 9 FINAL INTEGRITY** -> PASS (zero issues) -> Stage 10 (FAIL -> fix and re-verify)\n10. **Stage 10 HUMANIZE** -> user confirmation -> Stage 11\n11. **Stage 11 FINALIZE** -> MD + DOCX -> ask about LaTeX -> confirm -> PDF -> Stage 12\n12. **Stage 12 PROCESS SUMMARY** -> ask language version -> generate process record MD -> LaTeX -> PDF -> end\n\nSee `references/pipeline_state_machine.md` for complete state transition definitions.\n\n---\n\n## Adaptive Checkpoint System\n\n⚠️ **IRON RULE — Core rule: After each stage completion, the system must proactively prompt the user and wait for confirmation. The checkpoint presentation adapts based on context and user engagement.**\n\n### Checkpoint Types\n\n| Type | When Used | Content |\n|------|-----------|---------|\n| FULL | First checkpoint; after integrity boundaries; before finalization | Full deliverables list + decision dashboard + all options |\n| SLIM | After 2+ consecutive \"continue\" responses on non-critical stages | One-line status + auto-continue in 5 seconds |\n| MANDATORY | Integrity FAIL; Review decision; Stage 11 | Cannot be skipped; requires explicit user input |\n\n### Decision Dashboard (shown at FULL checkpoints)\n\n```\n━━━ Stage [X] [Name] Complete ━━━\n\nMetrics:\n- Word count: [N] (target: [T] +/-10%)    [OK/OVER/UNDER]\n- References: [N] (min: [M])              [OK/LOW]\n- Coverage: [N]/[T] sections drafted       [COMPLETE/PARTIAL]\n- Quality indicators: [score if available]\n\nDeliverables:\n- [Material 1]\n- [Material 2]\n\nFlagged: [any issues detected, or \"None\"]\n\nReady to proceed to Stage [Y]? You can also:\n1. View progress (say \"status\")\n2. Adjust settings\n3. Pause pipeline\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n```\n\n### Adaptive Rules\n\n1. **First checkpoint**: always FULL\n2. **After 2+ consecutive \"continue\" without review**: prompt user awareness (\"You've auto-continued [N] times. Want to review progress?\")\n3. **Integrity boundaries (Stage 4, 9)**: always MANDATORY\n4. **Review decisions (Stage 5, 6')**: always MANDATORY\n5. **Before finalization (Stage 11)**: always MANDATORY\n6. **All other stages**: start FULL, downgrade to SLIM if user says \"just continue\"\n\n### Checkpoint Rules\n\n1. ⚠️ **IRON RULE**: **Cannot auto-skip MANDATORY checkpoints**: Even if the previous stage result is perfect, explicit user input is required at MANDATORY checkpoints\n2. **User can adjust**: At FULL and MANDATORY checkpoints, users can modify the mode or settings for the next step\n3. **Pause-friendly**: Users can pause at any checkpoint and resume later\n4. **SLIM mode**: If the user says \"just continue\" or \"fully automatic,\" subsequent non-critical checkpoints switch to SLIM format (one-line status + auto-continue), but notifications are still sent\n5. **Awareness guard**: After 4+ consecutive auto-continues, the system inserts a FULL checkpoint regardless of stage type to ensure user remains engaged\n\n### Self-Check Questions (at every FULL checkpoint)\n\nBefore presenting the checkpoint to the user, the orchestrator asks itself:\n\n1. **Citation integrity**: Are there any unverified citations in the latest output?\n2. **Sycophantic concession**: Did the latest stage uncritically accept all feedback without pushback?\n3. **Quality trajectory**: Is the latest output ≥ the quality of the previous stage? If declining, PAUSE and flag.\n4. **Scope discipline**: Did the latest stage add content not requested by the user or the revision roadmap?\n5. **Completeness**: Are all required deliverables for this stage present?\n\nIf ANY answer raises concern, include it in the checkpoint presentation to the user.\n\n---\n\n## Agent Team (3 Agents)\n\n| # | Agent | Role | File |\n|---|-------|------|------|\n| 1 | `pipeline_orchestrator_agent` | Main orchestrator: detects stage, recommends mode, triggers skill, manages transitions | `agents/pipeline_orchestrator_agent.md` |\n| 2 | `state_tracker_agent` | State tracker: records completed stages, produced materials, revision loop count | `agents/state_tracker_agent.md` |\n| 3 | `integrity_verification_agent` | Integrity verifier: 100% reference/citation/data verification | `agents/integrity_verification_agent.md` |\n\n---\n\n## Orchestrator Workflow\n\n### Step 1: INTAKE & DETECTION\n\n```\npipeline_orchestrator_agent analyzes the user's input:\n\n1. What materials does the user have?\n   - No materials           --> Stage 2 (RESEARCH)\n   - Has research data      --> Stage 3 (WRITE)\n   - Has paper draft        --> Stage 4 (INTEGRITY)\n   - Has verified paper     --> Stage 5 (REVIEW)\n   - Has review comments    --> Stage 7 (REVISE)\n   - Has revised draft      --> Stage 6 (RE-REVIEW)\n   - Has final draft for formatting --> Stage 11 (FINALIZE)\n\n2. What is the user's goal?\n   - Full workflow (research to publication)\n   - Partial workflow (only certain stages needed)\n\n3. Determine entry point, confirm with user\n```\n\n### Step 2: MODE RECOMMENDATION\n\n```\nBased on entry point and user preferences, recommend modes for each stage:\n\nUser type determination:\n- Novice / wants guidance --> socratic (Stage 2) + plan (Stage 3) + guided (Stage 5)\n- Experienced / wants direct output --> full (Stage 2) + full (Stage 3) + full (Stage 5)\n- Time-limited --> quick (Stage 2) + full (Stage 3) + quick (Stage 5)\n\nExplain the differences between modes when recommending, letting the user choose\n```\n\n### Step 3: STAGE EXECUTION\n\n```\nCall the corresponding skill (does not do work itself, purely dispatching):\n\n1. Inform the user which Stage is about to begin\n2. Load the corresponding skill's SKILL.md\n3. Launch the skill with the recommended mode\n4. Monitor stage completion status\n\nAfter completion:\n1. Compile deliverables list\n2. Update pipeline state (call state_tracker_agent)\n3. [MANDATORY] Proactively prompt checkpoint, wait for user confirmation\n```\n\n### Step 4: TRANSITION\n\n```\nAfter user confirmation:\n\n1. Pass the previous stage's deliverables as input to the next stage\n2. Trigger handoff protocol (defined in each skill's SKILL.md):\n   - Stage 2  --> 2: deep-research handoff (RQ Brief + Bibliography + Synthesis)\n   - Stage 3  --> 4: Pass complete paper to integrity_verification_agent\n   - Stage 4 --> 3: Pass verified paper to reviewer\n   - Stage 5  --> 4: Pass Revision Roadmap to academic-paper revision mode\n   - Stage 7  --> 6: Pass revised draft and Response to Reviewers to reviewer\n   - Stage 6 --> 8: Pass new Revision Roadmap + R&R Traceability Matrix (Schema 11) to academic-paper revision mode\n   - Stage 7/8 --> 9: Pass revision-completed paper to integrity_verification_agent (final verification)\n   - Stage 9 --> 5: Pass verified final draft to format-convert mode\n3. Begin next stage\n```\n\n### Mid-Conversation Reinforcement Protocol\n\nAt every stage transition, the orchestrator MUST inject a brief core principles reminder. This prevents context rot in long conversations.\n\n**Template** (adapt to the upcoming stage):\n\n````\n--- STAGE TRANSITION: [Current] → [Next] ---\n\n🔄 Core Principles Reinforcement:\n1. [Most relevant IRON RULE for the next stage]\n2. [Most relevant Anti-Pattern to avoid in the next stage]\n3. Quality check: Is the output of [Current Stage] at least as good as [Previous Stage]? If not, PAUSE.\n\nCheckpoint: [MANDATORY/ADVISORY] — [What user needs to confirm]\n---\n````\n\n**Stage-specific reinforcement content**: See `references/reinforcement_content.md` for the full transition → reinforcement focus table.\n\n---\n\n## Integrity Review Protocol\n\nStage 4 (pre-review) and Stage 9 (post-revision) verification. 5-phase protocol: references → citation context → statistical data → originality → claims.\n\n⚠️ **IRON RULE**: Stage 9 must PASS with zero issues to proceed to Stage 11. Stage 9 verifies from scratch independently.\n\n⚠️ **IRON RULE (v3.2)**: Both Stage 4 and Stage 9 must also run the **AI Research Failure Mode Checklist** — a 7-mode taxonomy extending the citation hallucination checks into implementation bugs, hallucinated results, shortcut reliance, bug-as-insight, methodology fabrication, and pipeline-level frame-lock. If any of the 7 modes is `SUSPECTED`, or if Modes 1/3/5/6 are `INSUFFICIENT EVIDENCE`, the pipeline **blocks** and the user must acknowledge (confirm / override with reasoning / revise) before the pipeline proceeds. There is no `--no-block` escape hatch. Stage 12 PROCESS SUMMARY then reports the full failure-mode audit log as part of the AI Self-Reflection Report.\n\n> See `references/integrity_review_protocol.md` for the 5-phase citation/claim verification procedures.\n> See `references/ai_research_failure_modes.md` for the 7-mode AI research failure checklist and block/override logic.\n\n---\n\n## Two-Stage Review Protocol\n\nStage 5 (full review, 5 reviewers) → Revision Coaching → Stage 7 → Stage 6 (re-review) → optional Residual Coaching → Stage 8.\n\n> See `references/two_stage_review_protocol.md` for detailed stage flows and coaching dialogue limits.\n\n---\n\n## Mid-Entry Protocol\n\nUsers can enter from any stage. The orchestrator will:\n\n1. **Detect materials**: Analyze the content provided by the user to determine what is available\n2. **Identify gaps**: Check what prerequisite materials are needed for the target stage\n3. **Suggest backfilling**: If critical materials are missing, suggest whether to return to earlier stages\n4. **Direct entry**: If materials are sufficient, directly start the specified stage\n\n**Important: mid-entry cannot skip Stage 4**\n- If the user brings a paper and enters directly, go through Stage 4 (INTEGRITY) first before Stage 5 (REVIEW)\n- Only exception: User can provide a previous integrity verification report and content has not been modified\n\n---\n\n## External Review Protocol\n\nHandles external (human) reviewer feedback integration. 4-step workflow: Intake & Structuring → Strategic Revision Coaching → Revision & Response → Self-Verification.\n\n> See `references/external_review_protocol.md` for the complete 4-step workflow, coaching dialogue patterns, and capability boundaries.\n\n---\n\n## Progress Dashboard\n\nASCII dashboard shown at FULL checkpoints to display pipeline progress.\n\n> See `references/progress_dashboard_template.md` for the dashboard template.\n\n---\n\n## Revision Loop Management\n\n- Stage 5 (first review) -> Stage 7 (revision) -> Stage 6 (verification review) -> Stage 8 (re-revision, if needed) -> Stage 9 (final verification)\n- **Maximum 1 round of RE-REVISE** (Stage 8): If Stage 6 gives Major, enter Stage 8 for revision then proceed directly to Stage 9 (no return to review)\n- **Pipeline overrides academic-paper's max 2 revision rule**: In the pipeline, revisions are limited to Stage 7 + Stage 8 (one round each), replacing academic-paper's max 2 rounds rule\n- Mark unresolved issues as Acknowledged Limitations\n- Provide cumulative revision history (each round's decision, items addressed, unresolved items)\n\n### Early-Stopping Criterion (v3.2)\n\nAt the end of each revision round, if **delta < 3 points** on the 0-100 rubric AND **no P0 issues remain**, suggest stopping the revision loop (\"converged\"). User can override. Hard cap: 2 full revision loops (Stage 7 + Stage 8).\n\n### Budget Transparency (v3.2)\n\nAt pipeline start, estimate token cost based on paper length, mode, and cross-model toggle. Present estimate and ask for user confirmation before Stage 2 begins.\n\n---\n\n## Reproducibility\n\nEvery pipeline artifact is versioned, hashed, and auditable.\n\n> See `references/reproducibility_audit.md` for standardized workflow guarantees, audit trail format, and artifact tracking.\n\n---\n\n## Stage 12: Process Summary Protocol\n\nProduces the final process record: paper creation journey, collaboration quality evaluation (6 dimensions, 1-100), and AI self-reflection report.\n\n> See `references/process_summary_protocol.md` for full workflow, required content structure, scoring dimensions, and output specifications.\n\n---\n\n## Anti-Patterns\n\nExplicit prohibitions to prevent common failure modes:\n\n| # | Anti-Pattern | Why It Fails | Correct Behavior |\n|---|-------------|-------------|-----------------|\n| 1 | **Skipping integrity checks** | \"The paper looks fine, skip Stage 4, 9\" | Integrity checks are MANDATORY; they cannot be auto-skipped regardless of perceived quality |\n| 2 | **Orchestrator doing substantive work** | Pipeline orchestrator writes content or reviews the paper | Orchestrator only dispatches and coordinates; substantive work belongs to the sub-skills |\n| 3 | **Auto-advancing past MANDATORY checkpoints** | Moving to next stage without user confirmation at FULL checkpoints | MANDATORY checkpoints require explicit user input before proceeding |\n| 4 | **Quality degradation across stages** | Stage 7 revision is worse than Stage 3 draft because context window is exhausted | If Stage N output quality < Stage N-1, PAUSE and reload core principles before continuing |\n| 5 | **Silently dropping reviewer concerns** | Revision addresses 8 of 10 concerns and hopes nobody notices | The R&R tracking table must account for every concern with explicit status |\n| 6 | **Re-verifying only known issues at Stage 9** | Final integrity check only re-checks Stage 4 findings | Stage 9 must verify from scratch independently; revision may introduce new issues |\n| 7 | **Inflating Collaboration Quality scores** | Giving 90/100 to avoid awkward self-criticism | Honesty first: no inflation, no pleasantries; cite specific evidence for every score |\n| 8 | **Bypassing the Failure Mode Checklist block** (v3.2) | \"The 7-mode checklist is new, let's skip it this run\" | Stage 4, 9 Failure Mode Checklist is MANDATORY and BLOCKING; no `--no-block` flag exists; overrides require user reasoning recorded for Stage 12 |\n\n---\n\n## Quality Standards\n\n| Dimension | Requirement |\n|-----------|------------|\n| Stage detection | Correctly identify user's current stage and available materials |\n| Mode recommendation | Recommend appropriate mode based on user preferences and material status |\n| Material handoff | Stage-to-stage handoff materials are complete and correctly formatted |\n| State tracking | Pipeline state updated in real time; Progress Dashboard accurate |\n| **Mandatory checkpoint** | **User confirmation required after each stage completion** |\n| **Mandatory integrity check** | **Stage 4 and 9 cannot be skipped, must PASS** |\n| **Mandatory failure mode checklist** (v3.2) | **Stage 4 and 9 must run the 7-mode AI research failure checklist; suspected failures block; overrides require user reasoning** |\n| No overstepping | ⚠️ IRON RULE: Orchestrator does not perform substantive research/writing/reviewing, only dispatching |\n| No forcing | ⚠️ IRON RULE: User can pause or exit pipeline at any time (but cannot skip integrity checks) |\n| Reproducible | Same input follows the same workflow across different sessions |\n| **Convergence-aware stopping** (v3.2) | **If delta < 3 points AND no P0 issues, suggest stopping revision loop; user can override** |\n| **Budget transparency** (v3.2) | **Token cost estimate + user confirmation at pipeline start** |\n\n---\n\n## Error Recovery\n\n| Stage | Error | Handling |\n|-------|-------|---------|\n| Intake | Cannot determine entry point | Ask user what materials they have and their goal |\n| Stage 2 | deep-research not converging | Suggest mode switch (socratic -> full) or narrow scope |\n| Stage 3 | Missing research foundation | Suggest returning to Stage 2 to supplement research |\n| Stage 4 | Still FAIL after 3 correction rounds | List unverifiable items; user decides whether to continue |\n| Stage 5 | Review result is Reject | Provide options: major restructuring (Stage 3) or abandon |\n| Stage 7 | Revision incomplete on all items | List unaddressed items; ask whether to continue |\n| Stage 6 | Verification still has major issues | Enter Stage 8 for final revision |\n| Stage 8 | Issues remain after revision | Mark as Acknowledged Limitations; proceed to Stage 9 |\n| Stage 9 | Final verification FAIL | Fix and re-verify (max 3 rounds) |\n| Any | User leaves midway | Save pipeline state; can resume from breakpoint next time |\n| Any | Skill execution failure | Report error; suggest retry or skip |\n\n---\n\n## Agent File References\n\n| Agent | Definition File |\n|-------|----------------|\n| pipeline_orchestrator_agent | `agents/pipeline_orchestrator_agent.md` |\n| state_tracker_agent | `agents/state_tracker_agent.md` |\n| integrity_verification_agent | `agents/integrity_verification_agent.md` |\n\n---\n\n## Reference Files\n\n| Reference | Purpose |\n|-----------|---------|\n| `references/pipeline_state_machine.md` | Complete state machine definition: all legal transitions, preconditions, actions |\n| `references/plagiarism_detection_protocol.md` | Phase D originality verification protocol + self-plagiarism + AI text characteristics |\n| `references/mode_advisor.md` | Unified cross-skill decision tree: maps user intent to optimal skill + mode |\n| `references/claim_verification_protocol.md` | Phase E claim verification protocol: claim extraction, source tracing, cross-referencing, verdict taxonomy |\n| `references/ai_research_failure_modes.md` | 7-mode AI research failure checklist (Lu 2026), run at Stage 4 + 9 with blocking behaviour, reported at Stage 12 |\n| `references/team_collaboration_protocol.md` | Multi-person team coordination: role definitions, handoff protocol, version control, conflict resolution |\n| `references/integrity_review_protocol.md` | Stage 4 + 9 integrity verification: 5-phase protocol details |\n| `references/two_stage_review_protocol.md` | Two-stage review: Stage 5 full review + Stage 6 verification review |\n| `references/external_review_protocol.md` | External (human) reviewer feedback: 4-step intake/coaching/revision/verification |\n| `references/process_summary_protocol.md` | Stage 12: collaboration quality evaluation + AI self-reflection report |\n| `references/reproducibility_audit.md` | Standardized workflow guarantees + audit trail format |\n| `references/progress_dashboard_template.md` | ASCII progress dashboard template |\n| `references/reinforcement_content.md` | Stage-specific reinforcement focus table for transitions |\n| `references/changelog.md` | Full version history |\n| `shared/handoff_schemas.md` | Cross-skill data contracts: 9 schemas for all inter-stage handoff artifacts |\n\n---\n\n## Templates\n\n| Template | Purpose |\n|----------|---------|\n| `templates/pipeline_status_template.md` | Progress Dashboard output template |\n\n---\n\n## Examples\n\n| Example | Demonstrates |\n|---------|-------------|\n| `examples/full_pipeline_example.md` | Complete pipeline conversation log (Stage 2-5, with integrity + 2-stage review) |\n| `examples/mid_entry_example.md` | Mid-entry example starting from Stage 4 (existing paper -> integrity check -> review -> revision -> finalization) |\n\n---\n\n## Output Language\n\nFollows user language. Academic terminology retained in English.\n\n---\n\n## Integration with Other Skills\n\n```\nacademic-pipeline dispatches the following skills (does not do work itself):\n\nStage 2: deep-research\n  - socratic mode: Guided research exploration\n  - full mode: Complete research report\n  - quick mode: Quick research summary\n\nStage 3: academic-paper\n  - plan mode: Socratic chapter-by-chapter guidance\n  - full mode: Complete paper writing\n\nStage 4: integrity_verification_agent (Mode 1: pre-review)\nStage 9: integrity_verification_agent (Mode 2: final-check)\n\nStage 5: academic-paper-reviewer\n  - full mode: Complete 5-person review (EIC + R1/R2/R3 + Devil's Advocate)\n\nStage 6: academic-paper-reviewer\n  - re-review mode: Verification review (focused on revision responses)\n\nStage 7/8: academic-paper (revision mode)\nStage 11: academic-paper (format-convert mode)\n  - Step 1: Ask user which academic formatting style (APA 7.0 / Chicago / IEEE, etc.)\n  - Step 2: Auto-produce MD + DOCX\n  - Step 3: Produce LaTeX (using corresponding document class, e.g., apa7 class for APA 7.0)\n  - Step 4: After user confirms content is correct, tectonic compiles PDF (final version)\n  - Fonts: Times New Roman (English) + Source Han Serif TC VF (Chinese) + Courier New (monospace)\n  - ⚠️ IRON RULE: PDF must be compiled from LaTeX (HTML-to-PDF is prohibited)\n```\n\n---\n\n## Related Skills\n\n| Skill | Relationship |\n|-------|-------------|\n| `deep-research` | Dispatched (Stage 2 research phase) |\n| `academic-paper` | Dispatched (Stage 3 writing, Stage 7/8 revision, Stage 11 formatting) |\n| `academic-paper-reviewer` | Dispatched (Stage 5 first review, Stage 6 verification review) |\n\n---\n\n## Version Info\n\n| Item | Content |\n|------|---------|\n| Skill Version | 3.2 |\n| Last Updated | 2026-04-09 |\n| Maintainer | Cheng-I Wu |\n| Dependent Skills | deep-research v2.0+, academic-paper v2.0+, academic-paper-reviewer v1.1+ |\n| Role | Full academic research workflow orchestrator |\n\n---\n\n## Changelog\n\n> See `references/changelog.md` for full version history.\n\nFile v1.0.0:_meta.json\n\n{\n  \"ownerId\": \"kn79h8an2b3e37qcycc96jzpjs86qv7g\",\n  \"slug\": \"academic-pipeline\",\n  \"version\": \"1.0.0\",\n  \"publishedAt\": 1778726770557\n}\n\nFile v1.0.0:references/ai_research_failure_modes.md\n\n# AI Research Failure Mode Checklist\n\n**Status**: v3.2\n**Parent skill**: `academic-pipeline`\n**Used at**: Stage 2.5 INTEGRITY (blocking), Stage 4.5 FINAL INTEGRITY (blocking), Stage 6 PROCESS SUMMARY (reporting only)\n**Source**: Lu et al. (2026). Towards end-to-end automation of AI research. *Nature* 651, 914-919. doi:10.1038/s41586-026-10265-5 — Limitations section, Figure 2 (examples of failures in The AI Scientist's own accepted paper), Supplementary Information A.2.9 (debugging traces).\n\n---\n\n## Why this checklist exists\n\nLu et al. built the first autonomous AI research system to pass blind peer review (ICLR 2025 workshop). Their Limitations section enumerates the specific failure modes they observed — and most of them apply equally to human-in-the-loop AI research workflows like ARS.\n\nThese failures are dangerous because **they look like competent work**. A paper containing a hallucinated experimental result reads the same as a paper containing a real one. A shortcut-relying result reads the same as a genuine generalization. A methodology section describing experiments that were never actually run reads the same as a faithful account. The existing integrity verification catches citation hallucinations but is weak on the other failure modes.\n\nThe checklist exists to make these failures legible: **at Stage 2.5 and Stage 4.5, the integrity reviewer must explicitly rule out each of the 7 modes, or flag which are suspected and block the pipeline until the user acknowledges.**\n\nThis also extends the existing 5-type citation hallucination taxonomy (in `academic-paper-reviewer` references) into a broader 7-type AI research hallucination taxonomy. Citation hallucinations become mode 2 below.\n\n---\n\n## The 7 failure modes\n\n### Mode 1: Implementation bug passing AI self-review\n\n**What it is**: The analysis or experiment code has a bug (off-by-one, wrong variable, silent division-by-zero, type coercion, wrong flag) that produces numerically plausible but scientifically wrong results. The AI runs the code, looks at the output, sees nothing \"obviously\" wrong, and incorporates the result into the paper.\n\n**Lu 2026 example**: Supplementary A.2.9 traces show The AI Scientist repeatedly accepting experimental runs that had silent crashes or numerical instabilities because the top-level metric \"looked reasonable\". Figure 2 shows an ICLR reviewer catching one such issue in the accepted paper — the paper's main analysis depended on a setup that the code did not actually implement.\n\n**Detection questions at Stage 2.5**:\n- For every numerical result in the draft: does the user have a saved log, notebook, or script run that produced this number? If yes, was the exit code 0 and were there zero warnings? If no log is saved, flag.\n- Are any effect sizes suspiciously round (exactly 0.5, exactly 2x baseline, exactly zero variance across runs)? Suspiciously round numbers are a common signal of a constant leaking through a broken pipeline.\n- Do error bars / confidence intervals actually vary across conditions, or are they suspiciously identical?\n\n**Who catches it**: `methodology_reviewer_agent` at Stage 3 is downstream; the integrity gate at Stage 2.5 should ask the user directly.\n\n---\n\n### Mode 2: Hallucinated citation\n\n**What it is**: A reference that does not exist, is miscited (wrong year, wrong journal, wrong authors), or is attributed a finding it does not contain. This is the mode ARS already covers most thoroughly via the 5-type citation hallucination taxonomy in `academic-paper-reviewer/references/`. It is included here for completeness of the 7-mode taxonomy.\n\n**Lu 2026 example**: The AI Scientist pipeline includes a Semantic Scholar citation check to suppress this mode, acknowledging it as a primary failure class. PaperOrchestra (Song et al., 2026) extended this with a two-phase pipeline: web search discovery + sequential Semantic Scholar API verification (Levenshtein >= 0.70 title matching).\n\n**Detection (v3.3 update)**: Covered by the existing integrity verification, now strengthened with Semantic Scholar API batch verification (Phase A0 in `integrity_verification_agent`). See `deep-research/references/semantic_scholar_api_protocol.md` for the API protocol. The S2 API provides structured, machine-readable verification that catches fabricated DOIs (DOI_MISMATCH pattern) missed by manual WebSearch.\n\n**Who catches it**: `source_verification_agent` (Tier 0 S2 API + Tier 1 DOI + Tier 2 WebSearch) + `integrity_verification_agent` (Phase A0 + A1).\n\n---\n\n### Mode 3: Hallucinated experimental result\n\n**What it is**: A result that does not correspond to any actual experiment run. The AI writes \"we observed a 12% improvement\" when no run produced a 12% improvement — either it averaged differently than the paper claims, or it reported a number from a crashed run, or it invented the number to match the narrative.\n\n**Lu 2026 example**: Lu et al. specifically flag \"hallucinated experimental results\" as a Limitation, noting that automated reviewing struggles to detect them because the reviewer has no access to the underlying runs.\n\n**Detection questions at Stage 2.5**:\n- For every claim of the form \"X% improvement\" or \"Y% reduction\" or \"outperforms baseline by Z\": does the user have the raw numbers the paper's number was computed from?\n- Does the table in the draft match a saved CSV / tensor log / wandb run the user can point to?\n- Does the paper's \"we ran N seeds\" claim match the number of actual run directories the user has?\n\n**Who catches it**: integrity gate at 2.5. This is harder than citation checking because there's no external database to verify against — the verification is against the user's own experiment logs.\n\n---\n\n### Mode 4: Shortcut reliance\n\n**What it is**: The reported result is real but the model achieved it by exploiting a spurious feature rather than learning the intended generalization. A colour-biased MNIST model that gets 99% accuracy by reading the background colour, not the digit shape, is the canonical case.\n\n**Lu 2026 example**: Figure 2b shows this exact case — The AI Scientist proposed a method, tested it on colour-biased MNIST, got high accuracy, and wrote a paper claiming the method \"solved\" the task. A reviewer caught that the method was exploiting the colour shortcut, not learning the digit shape. The paper was revised.\n\n**Detection questions at Stage 2.5**:\n- Is there any controlled ablation that rules out the most obvious shortcut feature? If the paper tests on dataset D, has the author run on a D-variant where the shortcut feature is removed?\n- Does the paper's \"ablation studies\" section actually ablate the claimed mechanism, or does it ablate incidental hyperparameters?\n- Is the baseline strong enough that beating it requires the proposed mechanism, not just more compute?\n\n**Who catches it**: `devils_advocate_reviewer_agent` at Stage 3 is the natural home for this check, but it must be flagged at 2.5 so the user knows to prepare the ablation before Stage 3 arrives. Flag-only at 2.5, not block-only.\n\n---\n\n### Mode 5: Implementation bug reframed as novel insight\n\n**What it is**: The pipeline produces an unexpected result that is actually caused by a bug, but the narrative-writing stage reframes the unexpected behaviour as a novel finding. The paper claims \"we discovered that X behaves unexpectedly under condition Y\" when in reality X behaves the expected way and condition Y is being mis-implemented.\n\n**Lu 2026 example**: Lu et al. identify this as a compound failure mode — it requires Mode 1 (the bug) plus a writing-stage error in which the AI's narrative generator accepts the bug's output as real and builds a story around it. The paper reads *more* interesting than a bug-free version would have.\n\n**Detection questions at Stage 2.5**:\n- Does the draft contain any phrase like \"surprisingly,\" \"unexpectedly,\" \"counterintuitively,\" or \"contrary to our hypothesis\"? For each such claim, can the user point to a literature reference that would have predicted the opposite? If no literature cites the opposite, the \"surprise\" may not be a surprise — it may be a bug.\n- Did the surprising result appear on the first run, or only after many debugging iterations? First-run surprises are high-risk for this mode.\n- Has the user attempted to reproduce the surprising result from scratch in a fresh environment?\n\n**Who catches it**: integrity gate at 2.5. This is the most ARS-specific mode — it is the interaction of a bug (Mode 1) with narrative seduction.\n\n---\n\n### Mode 6: Methodology fabrication\n\n**What it is**: The Methods section describes experiments, hyperparameters, datasets, or procedures that were not actually what the pipeline ran. The AI writes a plausible-sounding Methods section based on what a reasonable version of the experiment *would* look like, drifting from what actually happened.\n\n**Lu 2026 example**: Lu et al. note that the writing stage of The AI Scientist sometimes produced Methods text that was disconnected from the actual hyperparameter log. They added a cross-check against the experiment run config to mitigate this.\n\n**Detection questions at Stage 2.5**:\n- Does every number in the Methods section (learning rate, batch size, epochs, dataset size, train/val split) appear in the user's actual run config / log?\n- Does the Methods section describe any preprocessing step the user cannot point to in their code?\n- Does the Methods section use the past tense where the actual pipeline didn't run?\n\n**Who catches it**: integrity gate at 2.5. This requires the user to provide the actual run config as an integrity input, not just the paper text.\n\n---\n\n### Mode 7: Frame-lock at early pipeline stage\n\n**What it is**: A wrong commitment made in early stages (research question framing, methodology choice, hyperparameter direction) that subsequent stages cannot back out of because they are structurally downstream of the commitment. The paper ends up well-executed but is answering the wrong question or using a fundamentally unsuitable method.\n\n**Lu 2026 example**: Figure 3a traces The AI Scientist's agentic tree search and shows that most failed papers failed at Stage 2 (hyperparameter tuning) — the agent committed to a direction early and could not recover. This is the same frame-lock pattern ARS's anti-sycophancy protocol targets for dialogue, but here it applies to pipeline decisions.\n\n**Detection questions at Stage 2.5**:\n- If the user could go back to Stage 1 knowing what they know now, would they change the research question or methodology?\n- Does the paper's Discussion section contain any phrase like \"in hindsight\" or \"we realized later\"? These are frame-lock tells.\n- Is the paper's contribution better explained by the chosen framing, or despite it?\n\n**Who catches it**: integrity gate at 2.5. If flagged, user is offered the option to return to Stage 1 or Stage 2 rather than proceeding to Stage 3.\n\n---\n\n## How the checklist runs at each stage\n\n### At Stage 2.5 INTEGRITY (first integrity gate)\n\nRun all 7 modes. For each mode, produce one of three outcomes:\n\n- **CLEAR**: integrity reviewer has evidence that the mode does not apply. Record the evidence briefly.\n- **SUSPECTED**: one or more detection questions returned a concerning answer. Must be surfaced to the user.\n- **INSUFFICIENT EVIDENCE**: integrity reviewer cannot rule the mode in or out without user input (e.g., needs experiment logs the user hasn't provided).\n\n**Block condition**: pipeline blocks if **any** mode is SUSPECTED, or if Modes 1, 3, 5, or 6 are INSUFFICIENT EVIDENCE (these four require user-provided logs to rule out and should not be silently skipped). Modes 2, 4, 7 INSUFFICIENT EVIDENCE can proceed with a warning and will be re-checked at 4.5.\n\n**User acknowledgement options at block**:\n- Confirm the flag — return to Stage 2 WRITE (or earlier) to fix\n- Override with reasoning — user explicitly states why the flag is a false positive, reasoning is recorded in the process log for Stage 6\n- Revise the specific passage and re-run the check\n\n### At Stage 4.5 FINAL INTEGRITY\n\nRe-run all 7 modes. Additional rule: any mode that was SUSPECTED at 2.5 must be resolved by 4.5 (CLEAR or user-Overridden-with-reasoning). If the same mode is still SUSPECTED at 4.5, the pipeline re-blocks and refuses to proceed to Finalize until the issue is addressed — no amount of revision loops can skip this.\n\n### At Stage 6 PROCESS SUMMARY (AI Self-Reflection Report)\n\nReport only, no blocking. The Self-Reflection Report includes a \"Failure Mode Audit Log\" section listing, for each of the 7 modes:\n- Final status at 4.5 (CLEAR / OVERRIDDEN)\n- History: was it ever SUSPECTED during the pipeline? At which stage? How was it resolved?\n- If OVERRIDDEN: the user's reasoning\n\nThis makes the failure-mode history part of the permanent process record, giving future readers (and the user themselves) visibility into what the AI-human collaboration had to defend against.\n\n---\n\n## Relationship to existing ARS checks\n\n| Existing check | Covers which modes |\n|---|---|\n| Citation hallucination taxonomy (5-type) | Mode 2 (fully) |\n| `source_verification_agent` | Mode 2 (cross-check) |\n| Existing Stage 2.5 integrity review | Mode 2, partial Mode 6 |\n| `devils_advocate_reviewer_agent` (Stage 3) | Mode 4, partial Mode 7 |\n| Anti-sycophancy protocol (v3.0) | Dialogue-level frame-lock, not pipeline-level Mode 7 |\n\nGap coverage provided by this checklist: **Modes 1, 3, 5, 6, and the pipeline-level aspect of Mode 7**. These are the modes that were not previously systematically checked.\n\n---\n\n## Open questions (for v3.3)\n\n- **False positive rate**: Modes 1, 5, and 6 require the user to supply experiment logs. If the user is writing a purely theoretical paper or a qualitative study, many of these detection questions don't apply. The checklist needs a paper-type pre-filter that turns off inapplicable modes based on the paper type detected by `field_analyst_agent`. v3.2 ships with all modes always-on; v3.3 should add the pre-filter.\n- **Override auditing**: if a user overrides a flag, is the reasoning ever reviewed? In v3.2 it goes into the Stage 6 record only. A stronger version would flag overrides for peer review during Stage 3 so that a reviewer can push back on the user's reasoning.\n\n---\n\n## References\n\n- Lu, C. et al. (2026). Towards end-to-end automation of AI research. *Nature* 651, 914-919. [doi:10.1038/s41586-026-10265-5](https://doi.org/10.1038/s41586-026-10265-5) — Limitations section, Figure 2, Supplementary Information A.2.9.\n- ARS `academic-paper-reviewer/references/` — existing 5-type citation hallucination taxonomy (Mode 2).\n- ARS `academic-pipeline/references/claim_verification_protocol.md` — existing integrity verification that this checklist extends.\n- ARS `academic-pipeline/references/integrity_review_protocol.md` — existing integrity review protocol that Stage 2.5 follows.\n- ARS `ROADMAP_v3.2.md` — v3.2 integration plan, item 2.\n\nFile v1.0.0:references/changelog.md\n\n# Changelog\n\n| Version | Date | Changes |\n|---------|------|---------|\n| 2.7 | 2026-03-27 | **Style Profile in Material Passport**: Pipeline orchestrator now carries optional Style Profile (Schema 10 in `shared/handoff_schemas.md`) through all stages. Produced by academic-paper intake Step 10 when user provides past writing samples. Consumed by draft_writer (Stage 2) and report_compiler (Stage 1) as soft writing voice guide. Does not affect integrity verification or review stages. Coordinates with deep-research v2.4 and academic-paper v2.5 |\n| 2.6 | 2026-03-08 | **Handoff Data Schema**: Enhanced `shared/handoff_schemas.md` with 9 comprehensive schemas (RQ Brief, Bibliography, Synthesis, Paper Draft, Integrity Report, Review Report, Revision Roadmap, Response to Reviewers, Material Passport) with full field definitions, type constraints, and validation rules; orchestrator validates output against schemas before each transition. **Adaptive Checkpoint System**: Replaced static checkpoint template with 3-tier system (FULL/SLIM/MANDATORY) based on stage criticality and user engagement; FULL checkpoints include decision dashboard with metrics; SLIM auto-continues for experienced users; MANDATORY cannot be bypassed at integrity/review/finalization boundaries; awareness guard after 4+ auto-continues. **Mode Advisor**: New `references/mode_advisor.md` with unified cross-skill decision tree, common misconceptions table, user archetype recommendations, decision flowchart, and anti-patterns guide. **Team Collaboration Protocol**: New `references/team_collaboration_protocol.md` with 5 role definitions, per-transition handoff procedures, git branching/tagging strategy, conflict resolution matrix, and communication templates; state tracker extended with `assigned_to`, `approval_gate`, `team_notes` per stage and `schema_validation_log`. **Phase E Claim Verification**: New `references/claim_verification_protocol.md` with E1 claim extraction, E2 source tracing, E3 cross-referencing; verdict taxonomy (VERIFIED / MINOR_DISTORTION / MAJOR_DISTORTION / UNVERIFIABLE / UNVERIFIABLE_ACCESS); severity mapping (MAJOR_DISTORTION -> SERIOUS, UNVERIFIABLE -> SERIOUS, MINOR_DISTORTION -> MINOR, UNVERIFIABLE_ACCESS -> MEDIUM); integrated into integrity_verification_agent Mode 1 (30% spot-check) and Mode 2 (100%); pass/fail criteria updated to include Phase E verdicts. **Mid-Entry Material Passport Check**: Pipeline orchestrator now validates Material Passport on mid-entry; decision tree checks verification_status, freshness (< 24 hours), and content modification (version_label comparison); offers skip/spot-check/full re-verify options for Stage 2.5 when passport is valid; passport freshness validation rules added to `shared/handoff_schemas.md` |\n| 2.5 | 2026-03-08 | External Review Protocol: structured intake of real journal reviewer feedback (text/PDF/DOCX); 4-step workflow (parse -> strategic coaching -> revise + Response to Reviewers -> completeness check); differentiated behavior from internal simulated review (no default \"accept all\", risk assessment per comment, user confirmation of parsed items); explicit capability boundaries (AI verification ≠ reviewer satisfaction) |\n| 2.4 | 2026-03-08 | Stage 6 PROCESS SUMMARY: post-pipeline paper creation process record; asks user preferred language (zh/en/both); generates structured MD summarizing full human-AI collaboration history with user quotes, key decisions, iteration details, and lessons learned; mandatory final chapter: **Collaboration Quality Evaluation** (6 dimensions scored 1-100, bar chart visualization, What Worked Well / Missed Opportunities / Recommendations / Human vs AI Value-Add / Claude's Self-Reflection); compiles to PDF via LaTeX + tectonic; outputs `paper_creation_process_zh.pdf` + `paper_creation_process_en.pdf` |\n| 2.3 | 2026-03-08 | Stage 5 FINALIZE: mandatory formatting style prompt (APA 7.0 / Chicago / IEEE); PDF must compile from LaTeX via tectonic (no HTML-to-PDF); APA 7.0 uses `apa7` document class (`man` mode) with XeCJK for bilingual support; font stack: Times New Roman + Source Han Serif TC VF + Courier New |\n| 2.2 | 2026-03-05 | Checkpoint confirmation semantics (6 user commands with precise actions); mode switching rules (safe/dangerous/prohibited matrix); skill failure fallback matrix (per-stage degradation strategies); state ownership protocol (single source of truth with write access control); material version control (versioned artifacts with audit trail); cross-skill reference to `shared/handoff_schemas.md` |\n| 2.1 | 2026-03 | Added plagiarism detection protocol (Phase D); enhanced integrity_verification_agent with originality verification (D1 WebSearch, D2 self-plagiarism); updated both verification modes |\n| 2.0 | 2026-02 | Added Stage 2.5/4.5 integrity checks, two-stage review, mandatory checkpoints, Devil's Advocate, reproducibility guarantees, integrity_verification_agent |\n| 1.0 | 2026-02 | Initial version: 5+1 stage pipeline |\n\nFile v1.0.0:references/claim_verification_protocol.md\n\n# Claim Verification Protocol (Phase E)\n\n## Purpose\nVerifies that quantitative and factual claims in the paper are accurately supported by their cited sources. Phase A-D verify that references exist and are original; Phase E verifies that claims derived from those references are truthful.\n\n## Scope\n- All numerical claims (percentages, counts, effect sizes, p-values)\n- All categorical assertions (\"X is the largest...\", \"Y was the first to...\")\n- All trend claims (\"increasing\", \"declining\", \"stable\")\n- All causal claims (\"X causes Y\", \"X leads to Y\")\n\n## E1: Claim Extraction\n- Scan the paper for all quantitative/factual claims\n- For each claim, record: claim text, cited source(s), paper section, page/line\n- Expected output: Claim Registry table\n\n## E2: Source Tracing\n- For each claim, locate the specific passage in the cited source that supports it\n- Use WebSearch + DOI lookup to find the original source\n- If source is behind paywall, note as UNVERIFIABLE_ACCESS\n\n## E3: Cross-Referencing\n- Compare claim text vs source text\n- Check: exact numbers, date ranges, population descriptions, methodology descriptions\n- Flag any discrepancies\n\n## Verdict Taxonomy\n\n| Verdict | Definition | Severity | Example |\n|---------|-----------|----------|---------|\n| VERIFIED | Claim matches source exactly or within rounding tolerance | None | Paper: \"15.2%\"; Source: \"15.2%\" |\n| MINOR_DISTORTION | Claim paraphrases source but meaning is preserved | MINOR | Paper: \"about 15%\"; Source: \"15.2%\" |\n| MAJOR_DISTORTION | Claim oversimplifies, exaggerates, or misrepresents source | SERIOUS | Paper: \"declined sharply\"; Source: \"declined by 2.1%\" |\n| UNVERIFIABLE | Source doesn't contain the claimed information | SERIOUS | Paper cites Smith (2020) for a claim, but Smith (2020) doesn't discuss this topic |\n| UNVERIFIABLE_ACCESS | Source exists but full text not accessible for verification | MEDIUM | Paywalled journal article |\n\n## Sampling Strategy\n- Mode 1 (pre-review): 30% random sample of claims (minimum 10 claims)\n- Mode 2 (final-check): 100% of claims\n\n## Output Format\n\n### Claim Verification Report\n| # | Claim | Source | Section | Verdict | Detail |\n|---|-------|-------|---------|---------|--------|\n| 1 | [claim text] | [source] | [section] | VERIFIED | Exact match |\n| 2 | [claim text] | [source] | [section] | MAJOR_DISTORTION | Paper says X, source says Y |\n\n### Summary\n- Total claims checked: [N]\n- VERIFIED: [N]\n- MINOR_DISTORTION: [N]\n- MAJOR_DISTORTION: [N] (must be 0 for PASS)\n- UNVERIFIABLE: [N] (must be 0 for PASS)\n- UNVERIFIABLE_ACCESS: [N] (noted but does not block PASS)\n\n## Pass/Fail Criteria\n- PASS: Zero MAJOR_DISTORTION + Zero UNVERIFIABLE\n- FAIL: Any MAJOR_DISTORTION or UNVERIFIABLE\n- PASS_WITH_NOTES: Only MINOR_DISTORTION and/or UNVERIFIABLE_ACCESS\n\nFile v1.0.0:references/external_review_protocol.md\n\n# External Review Protocol (Added in v2.5)\n\n**Scenario**: The user submitted to a journal and received feedback from real human reviewers, bringing those comments into the pipeline.\n\n**Trigger**: User says \"I received reviewer comments,\" \"reviewer comments,\" \"revise and resubmit,\" etc.\n\n## Differences from Internal Review\n\n| Aspect | Internal Review (Stage 3 simulation) | External Review (real journal) |\n|--------|-------------------------------------|-------------------------------|\n| Source of review comments | Pipeline's AI reviewers | Journal's human reviewers |\n| Comment format | Structured (Revision Roadmap) | Unstructured (free text, PDF, email) |\n| Comment quality | Consistent, predictable | Variable quality, may be vague or contradictory |\n| Revision strategy | Can accept wholesale | Need to judge which to accept/reject/negotiate |\n| Acceptance criteria | AI re-review suffices | Ultimately decided by human reviewers |\n\n## Step 1: Intake and Structuring\n\n```\n1. Receive reviewer comments (supported formats):\n   - Directly pasted text\n   - Provide PDF/DOCX file path\n   - Copy from journal system review letter\n\n2. Parse into structured list:\n   For each comment, extract:\n   - Reviewer number (Reviewer 1/2/3 or R1/R2/R3)\n   - Comment type: Major / Minor / Editorial / Positive\n   - Core request (one-sentence summary)\n   - Original text quote\n   - Paper section involved\n\n3. Produce External Review Summary:\n   +----------------------------------------+\n   | External Review Summary                |\n   +----------------------------------------+\n   | Journal: [journal name]                |\n   | Decision: [R&R / Major / Minor]        |\n   | Reviewers: [N]                         |\n   | Total comments: [N]                    |\n   |   Major: [n]  Minor: [n]  Editorial: [n]|\n   +----------------------------------------+\n\n4. Confirm parsing results with user:\n   \"I organized the reviewer comments into [N] items. Here is the summary — please confirm nothing was missed or misinterpreted.\"\n```\n\n## Step 2: Strategic Revision Coaching (External Revision Coaching)\n\nUnlike the Socratic coaching for internal review, external review coaching focuses more on **strategic judgment**:\n\n```\nFor each Major comment, guide the user to think through:\n\n1. Understanding layer\n   \"What is this reviewer's core concern? Is it about methodology, theory, or presentation?\"\n\n2. Judgment layer\n   \"Do you agree with this criticism?\"\n   - Agree -> \"How do you plan to revise?\"\n   - Partially agree -> \"Which parts do you agree with and which not? What is your basis for disagreement?\"\n   - Disagree -> \"What is your rebuttal argument? Can you support it with literature or data?\"\n\n3. Strategy layer\n   \"How will you phrase this in the response letter?\"\n   - Accept revision: Show specifically what was changed and where\n   - Partially accept: Explain the accepted parts + reasons for non-acceptance (must be persuasive)\n   - Reject: Provide sufficient scholarly rationale (literature, data, methodological argumentation)\n\n4. Risk assessment\n   \"If you reject this suggestion, what might the reviewer's reaction be? Is it worth the risk?\"\n```\n\n**Key principles**:\n- **Do not default to \"accept all\"**: Real reviewer comments are not always correct — some may be based on misunderstanding or school-of-thought bias\n- **Encourage user to inject context**: \"What school of thought do you think this reviewer might come from? What context might they not be aware of?\"\n- **User can say \"just fix it for me\" to skip**: But when skipping strategic discussion, AI defaults to accepting all comments (conservative strategy)\n- **Maximum 8 rounds of dialogue**, but at least 1 round per Major comment\n\n## Step 3: Revision and Response to Reviewers\n\n```\nProduce two documents:\n\n1. Revised draft\n   - Track all modification locations (additions/deletions/rewrites)\n   - Revision content consistent with Response to Reviewers\n\n2. Response to Reviewers letter\n   Format (point-by-point response):\n   +------------------------------------+\n   | Reviewer [N], Comment [M]:         |\n   |                                    |\n   | [Original comment quote]           |\n   |                                    |\n   | Response:                          |\n   | [Response explanation]             |\n   |                                    |\n   | Changes made:                      |\n   | [Specific modification location    |\n   |  and content]                      |\n   | (or: We respectfully disagree      |\n   |  because... [rationale])           |\n   +------------------------------------+\n```\n\n## Step 4: Self-Verification (Completeness Check)\n\n```\nStage 3' behavior adjustments in external review mode:\n\n1. Point-by-point comparison of External Review Summary with Response to Reviewers:\n   - Does every comment have a response? (completeness)\n   - Is each response consistent with actual changes? (consistency)\n   - Were the places claimed as \"modified\" actually changed? (truthfulness)\n\n2. New citation verification:\n   - New references added during revision enter Stage 4.5 integrity verification\n\n3. Things NOT done (different from internal review):\n   - Do not reassess paper quality (that is the human reviewers' job)\n   - Do not issue a new Editorial Decision\n   - Do not raise new revision requests\n```\n\n## Honest Capability Boundaries\n\n1. **AI verification does not equal human reviewer satisfaction**: Stage 3' can confirm revisions are \"complete and consistent,\" but cannot predict whether human reviewers will accept your responses. Reviewers may have unstated expectations, school-of-thought preferences, or methodological insistence\n2. **Unstructured comments may not parse perfectly**: Some reviewers write vaguely (e.g., \"the methodology needs more work\"), and AI will do its best to parse but may miss implied intentions. After parsing, **user confirmation is mandatory**\n3. **AI cannot make scholarly judgments for you**: \"Should I accept Reviewer 2's suggestion?\" is your decision. AI can provide an analytical framework, but final judgment rests with the researcher\n4. **Cross-cultural review convention differences**: Response conventions differ across journals/academic circles (some require extreme deference, others accept direct rebuttal). AI defaults to neutral academic tone; the user can request adjustments\n\nFile v1.0.0:references/integrity_review_protocol.md\n\n# Integrity Review Protocol (Added in v2.0)\n\n## Stage 2.5: First Integrity Check (Pre-Review Integrity)\n\n**Trigger**: After Stage 2 (WRITE) completion, before Stage 3 (REVIEW)\n**Purpose**: Ensure all references and data are not fabricated or erroneous before submission for review\n\n```\nExecution steps:\n1. integrity_verification_agent executes Mode 1 (initial verification) on the paper\n2. Verification scope:\n   - Phase A: 100% reference existence + bibliographic accuracy + ghost citations\n   - Phase B: >= 30% citation context spot-check\n   - Phase C: 100% statistical data verification\n   - Phase D: >= 30% originality spot-check + self-plagiarism check\n   - Phase E: 30% claim verification spot-check (minimum 10 claims)\n3. Result handling:\n   - PASS -> checkpoint -> Stage 3\n   - FAIL -> produce correction list -> fix item by item -> re-verify corrected items\n   - PASS after corrections -> checkpoint -> Stage 3\n   - Still FAIL after 3 rounds -> notify user, list unverifiable items\n```\n\n## Stage 4.5: Final Integrity Check (Post-Revision Final Check)\n\n**Trigger**: After Stage 4' (RE-REVISE) or Stage 3' (RE-REVIEW, Accept) completion, before Stage 5 (FINALIZE)\n**Purpose**: Confirm the revised paper is 100% correct and ready for publication\n\n```\nExecution steps:\n1. integrity_verification_agent executes Mode 2 (final verification) on the revised draft\n2. Verification scope:\n   - Phase A: 100% reference verification (including those added during revision)\n   - Phase B: 100% citation context verification (not spot-check, full check)\n   - Phase C: 100% statistical data verification\n   - Phase D: >= 50% originality spot-check (100% for newly added/modified paragraphs)\n   - Phase E: 100% claim verification (zero MAJOR_DISTORTION + zero UNVERIFIABLE required)\n3. Special check: Compare with Stage 2.5 results to confirm all previous issues are resolved\n4. Result handling:\n   - PASS (zero issues) -> checkpoint -> Stage 5\n   - FAIL -> fix -> re-verify -> PASS -> Stage 5\n5. ⚠️ **IRON RULE**: Must PASS with zero issues to proceed to Stage 5\n```\n\n## Score Trajectory Tracking (v3.3)\n\nReference: `academic-pipeline/references/score_trajectory_protocol.md`\n\nAt Stage 3' (RE-REVIEW), the `pipeline_orchestrator_agent` tracks per-dimension score deltas and triggers a MANDATORY checkpoint on regressions. Results stored in Integrity Report `score_trajectory` field (Schema 5).\n\nFile v1.0.0:references/mode_advisor.md\n\n# Mode Advisor — Unified Cross-Skill Decision Tree\n\n## Purpose\n\nHelps users (and the pipeline orchestrator) select the right skill and mode for their current situation. Eliminates the most common routing mistakes by mapping user intent to the optimal entry point.\n\n---\n\n## Quick Decision Matrix\n\n| What do you want? | How far along? | Time? | Skill + Mode |\n|-------------------|---------------|-------|--------------|\n| Explore a topic | Starting fresh | 30 min | deep-research quick |\n| Explore a topic | Starting fresh | 2+ hr | deep-research full |\n| Think through a research idea | Have vague idea | Any | deep-research socratic |\n| Systematic review | Have clear PICO | 3+ hr | deep-research systematic-review |\n| Verify claims | Have specific claims | 30 min | deep-research fact-check |\n| Write a paper | Have research done | 2+ hr | academic-paper full |\n| Plan a paper step by step | Have RQ, need structure | 1+ hr | academic-paper plan |\n| Fix citations | Have draft | 30 min | academic-paper citation-check |\n| Convert format | Have final draft | 15 min | academic-paper format-convert |\n| Review a paper | Have paper to evaluate | 1 hr | academic-paper-reviewer full |\n| Check revision quality | Have revised draft | 30 min | academic-paper-reviewer re-review |\n| Full pipeline (zero to publication) | Starting fresh | 5+ hr | academic-pipeline |\n| Handle real reviewer feedback | Have review comments | 1+ hr | academic-pipeline (Stage 4 entry) |\n\n---\n\n## Common Misconceptions\n\n| User Says | They Probably Need | Why |\n|-----------|-------------------|-----|\n| \"Write me a paper on X\" | deep-research first, THEN academic-paper | Writing without research produces shallow papers with unsupported claims |\n| \"Review my paper\" (but no draft exists) | academic-paper plan mode | They need to write first, not review |\n| \"Check my citations\" (but paper isn't done) | academic-paper full mode | Finish writing first, then check citations as a separate pass |\n| \"I need a systematic review\" | deep-research systematic-review mode | NOT academic-paper lit-review structure (different methodology: PRISMA vs narrative) |\n| \"Just give me a quick paper\" | deep-research quick + academic-paper full | Quick research is fine, but paper writing still needs the full mode for quality |\n| \"Format my paper as APA\" | academic-paper format-convert mode | Not a rewrite; purely formatting transformation |\n| \"I got reviewer comments\" | academic-pipeline Stage 4 entry (External Review) | Needs structured intake + strategic coaching, not just \"fix what they said\" |\n\n---\n\n## User Archetype Recommendations\n\n| Archetype | Recommended Workflow | Rationale |\n|-----------|---------------------|-----------|\n| Graduate student (first paper) | deep-research socratic -> academic-paper plan -> full pipeline | Socratic mode builds research thinking; plan mode structures the paper incrementally; pipeline ensures quality gates |\n| Experienced researcher (submission prep) | academic-pipeline (full, from Stage 1 or mid-entry) | Knows what they want; benefits from the automated quality assurance and integrity checks |\n| Advisor reviewing student work | academic-paper-reviewer full | Provides structured multi-perspective feedback the advisor can use in mentoring |\n| Quick literature scan | deep-research quick or lit-review | Fast turnaround; no need for full pipeline overhead |\n| Journal revision response | academic-pipeline (Stage 4 entry with review comments) | External Review Protocol handles real reviewer feedback with strategic coaching |\n| Conference paper (short deadline) | deep-research quick -> academic-paper full (conference type) | Compressed timeline; quick research + full writing with conference structure |\n| Thesis chapter | deep-research full -> academic-paper full | Each chapter treated as a standalone paper; full depth needed |\n| Policy brief | deep-research quick -> academic-paper full (policy_brief type) | Evidence-based but concise; quick research sufficient for policy scope |\n\n---\n\n## Skill Capability Boundaries\n\nUnderstanding what each skill can and cannot do prevents misrouting:\n\n| Skill | Can Do | Cannot Do |\n|-------|--------|-----------|\n| deep-research | Literature search, synthesis, RQ refinement, fact-checking | Write papers, review papers, format documents |\n| academic-paper | Write papers, revise papers, format documents, check citations | Conduct original research, review papers (as reviewer), verify integrity |\n| academic-paper-reviewer | Review papers (5-person panel), re-review revisions | Write papers, conduct research, fix issues (only identifies them) |\n| academic-pipeline | Orchestrate all stages, manage transitions, track state | Perform any substantive work (purely dispatching and coordinating) |\n| integrity_verification_agent | Verify references, citations, data, originality | Fix issues (only identifies them), review paper quality |\n\n---\n\n## Decision Flowchart\n\n```\nSTART: What does the user want?\n  |\n  +--> \"I want to research/explore/investigate\"\n  |      |\n  |      +--> Have specific claims to verify? --> deep-research fact-check\n  |      +--> Have clear PICO/systematic question? --> deep-research systematic-review\n  |      +--> Want guided exploration? --> deep-research socratic\n  |      +--> Want direct results, have time? --> deep-research full\n  |      +--> Want direct results, short on time? --> deep-research quick\n  |\n  +--> \"I want to write a paper\"\n  |      |\n  |      +--> Have research/literature ready? --> academic-paper (plan or full)\n  |      +--> No research done yet? --> deep-research FIRST, then academic-paper\n  |      +--> Want full quality assurance? --> academic-pipeline (from Stage 1)\n  |\n  +--> \"I want someone to review my paper\"\n  |      |\n  |      +--> Have a complete draft? --> academic-paper-reviewer full\n  |      +--> Want integrity check + review? --> academic-pipeline (Stage 2.5 entry)\n  |      +--> No draft yet? --> academic-paper first\n  |\n  +--> \"I need to revise based on feedback\"\n  |      |\n  |      +--> From AI reviewers (pipeline)? --> Continue pipeline (Stage 4)\n  |      +--> From real journal reviewers? --> academic-pipeline Stage 4 entry (External Review)\n  |\n  +--> \"I want the full treatment (research to publication)\"\n         |\n         +--> academic-pipeline (Stage 1 entry)\n```\n\n---\n\n## Pipeline Stage Entry Points\n\nFor users entering the pipeline mid-stream, this table clarifies what materials are needed:\n\n| Entry Point | Required Materials | What Gets Skipped | Integrity Implications |\n|------------|-------------------|-------------------|----------------------|\n| Stage 1 (RESEARCH) | None | Nothing | Full pipeline |\n| Stage 2 (WRITE) | RQ Brief + Bibliography | Stage 1 | Full pipeline from Stage 2 |\n| Stage 2.5 (INTEGRITY) | Paper draft | Stages 1-2 | Integrity check runs on provided draft |\n| Stage 3 (REVIEW) | Verified paper + integrity report | Stages 1-2.5 | User must provide integrity evidence |\n| Stage 4 (REVISE) | Paper + review comments | Stages 1-3 | Pipeline runs Stage 4 -> 3' -> 4' -> 4.5 -> 5 |\n| Stage 5 (FINALIZE) | Paper + integrity pass report | Stages 1-4.5 | Must show Stage 4.5 passed |\n\n---\n\n## Anti-Patterns\n\nThese are common workflow mistakes to avoid:\n\n| Anti-Pattern | Problem | Correct Approach |\n|-------------|---------|-----------------|\n| Skipping research | Paper lacks evidence depth | Always do at least deep-research quick |\n| Writing then researching | Confirmation bias in source selection | Research first, write second |\n| Reviewing before integrity check | Wasted review effort on fabricated citations | Always Stage 2.5 before Stage 3 |\n| Accepting all reviewer comments blindly | May introduce inconsistencies or weaken valid arguments | Use External Review Protocol's strategic coaching |\n| Running pipeline for a 1-page abstract | Overhead far exceeds benefit | Use academic-paper full directly |\n| Using fact-check mode for literature review | Different purpose and methodology | Use deep-research full or systematic-review |\n\nFile v1.0.0:references/pipeline_state_machine.md\n\n# Pipeline State Machine v2.0 — Complete Definition\n\nThis document defines all legal states, transition conditions, transition actions, and exception handling for academic-pipeline v2.0.\n\n---\n\n## State Definitions\n\n### Stage States\n\n| State | Description |\n|-------|------------|\n| `pending` | Not yet started, waiting for prerequisite stage to complete |\n| `in_progress` | Currently executing |\n| `completed` | Completed, deliverables recorded |\n| `skipped` | User chose to skip (only for non-mandatory stages) |\n| `blocked` | Preconditions not met (e.g., integrity check FAIL) |\n\n### Pipeline Global States\n\n| State | Description |\n|-------|------------|\n| `initializing` | Detecting entry point and materials |\n| `running` | Pipeline executing (at least one stage is in_progress) |\n| `awaiting_confirmation` | Stage complete, waiting for user to confirm checkpoint |\n| `paused` | User paused, can resume at any time |\n| `completed` | All required stages complete, final paper produced |\n| `aborted` | User abandoned (e.g., chose to abandon after Reject) |\n\n---\n\n## State Transition Diagram (ASCII)\n\n```\n                        +-------------+\n                        | INITIALIZING|\n                        +------+------+\n                               |\n                    [Detect entry point & materials]\n                               |\n         +----------+----------+----------+----------+\n         |          |          |          |          |\n         v          v          v          v          v\n    +--------+ +--------+ +--------+ +--------+ +--------+\n    |Stage 1 | |Stage 2 | |Stg 2.5 | |Stage 3 | |Stage 4 |\n    |RESEARCH| | WRITE  | |INTEGRIT| | REVIEW | | REVISE |\n    +---+----+ +---+----+ +---+----+ +---+----+ +---+----+\n        |          |          |          |          |\n   [checkpoint]   [checkpoint]   |     [checkpoint]  |\n        |          |          |          |          |\n        v          v          v          v          v\n   +--------+ +--------+ +---+----+    |          |\n   |Stage 2 | |Stg 2.5 | |PASS?   |    |          |\n   | WRITE  | |INTEGRIT| +---+----+    |          |\n   +---+----+ +---+----+     |         |          |\n                         +----+----+    |          |\n                         |         |    |          |\n                        Yes       No    |          |\n                         |     [Fix]    |          |\n                         |   [Re-verify]|          |\n                    [checkpoint]   |    |          |\n                         |         |    |          |\n                         v         |    |          |\n                    +--------+     |    |          |\n                    |Stage 3 | <---+    |          |\n                    | REVIEW |          |          |\n                    +---+----+          |          |\n                        |               |          |\n                   [DECISION]           |          |\n                        |               |          |\n              +---------+---------+     |          |\n              |         |         |     |          |\n            Accept    Minor     Major   |          |\n              |       Revision  Revision|          |\n              |         |         |     |          |\n              |    [checkpoint]  [checkpoint]      |\n              |         |         |     |          |\n              |         v         v     |          |\n              |    +--------+ +--------+|          |\n              |    |Stage 4 | |Stage 4 ||          |\n              |    | REVISE | | REVISE ||          |\n              |    +---+----+ +---+----+|          |\n              |        |          |     |          |\n              |   [checkpoint]   [checkpoint]      |\n              |        |          |     |          |\n              |        v          v     |          |\n              |    +--------+ +--------+           |\n              |    |Stg 3'  | |Stg 3'  |           |\n              |    |RE-REV. | |RE-REV. |           |\n              |    +---+----+ +---+----+           |\n              |        |          |                 |\n              |   [DECISION]  [DECISION]            |\n              |        |          |                 |\n              |     Accept      Major               |\n              |     /Minor        |                 |\n              |        |     [checkpoint]           |\n              |        |          |                 |\n              |        |          v                 |\n              |        |     +--------+             |\n              |        |     |Stg 4'  |             |\n              |        |     |RE-REVIS|             |\n              |        |     +---+----+             |\n              |        |          |                 |\n              |   [checkpoint]  [checkpoint]        |\n              |        |          |                 |\n              v        v          v                 |\n         +----+--------+----------+-----+           |\n         |     Stage 4.5                |           |\n         |   FINAL INTEGRITY            |           |\n         +----------+------------------+           |\n                    |                               |\n               [PASS? Zero issues]                  |\n                    |                               |\n              +-----+-----+                         |\n              |           |                         |\n             Yes         No                         |\n              |        [Fix]                         |\n              |      [Re-verify]                     |\n         [checkpoint]     |                         |\n              |           |                         |\n              v           |                         |\n         +--------+       |                         |\n         |Stage 5 | <-----+                         |\n         |FINALIZE|                                 |\n         +---+----+                                 |\n             |                                      |\n             v                                      |\n         +-------+                                  |\n         |  END  |                                  |\n         +-------+                                  |\n```\n\n---\n\n## Legal State Transitions\n\n### Normal Flow Transitions\n\n| From | To | Precondition | Action |\n|------|----|-------------|--------|\n| INIT | Stage 1 | User confirms starting from Stage 1 | Detect mode preference, launch deep-research |\n| INIT | Stage 2 | User has research materials, confirms skipping Stage 1 | Detect materials, launch academic-paper |\n| INIT | Stage 2.5 | User has complete paper | Launch integrity_verification_agent |\n| INIT | Stage 3 | User has verified paper + integrity report | Confirm paper language/domain, launch reviewer |\n| INIT | Stage 4 | User has review comments | Confirm paper + review comments, launch revision |\n| INIT | Stage 5 | User has final draft for format conversion | Confirm format requirements, launch format-convert |\n| Stage 1 | **checkpoint** | Stage 1 completed | Wait for user confirmation |\n| checkpoint | Stage 2 | User confirms | handoff RQ Brief + Bibliography + Synthesis |\n| Stage 2 | **checkpoint** | Stage 2 completed, Paper Draft produced | Wait for user confirmation |\n| checkpoint | Stage 2.5 | User confirms | Pass Paper Draft to integrity agent |\n| Stage 2.5 | **checkpoint** | PASS | Wait for user confirmation |\n| Stage 2.5 | Stage 2.5 (retry) | FAIL | Fix issues, re-verify (max 3 rounds) |\n| checkpoint | Stage 3 | User confirms | Pass verified paper to reviewer |\n| Stage 3 | **checkpoint** | Decision produced | Wait for user confirmation |\n| checkpoint | Stage 4 | Decision = Minor/Major, user confirms | Pass Revision Roadmap |\n| checkpoint | Stage 4.5 | Decision = Accept, user confirms | Skip revision, go directly to final verification |\n| Stage 4 | **checkpoint** | Stage 4 completed | Wait for user confirmation |\n| checkpoint | Stage 3' | User confirms | Pass Revised Draft + Response to Reviewers |\n| Stage 3' | **checkpoint** | Decision produced | Wait for user confirmation |\n| checkpoint | Stage 4.5 | Decision = Accept/Minor, user confirms | Pass final draft to final verification |\n| checkpoint | Stage 4' | Decision = Major, user confirms | Pass new Revision Roadmap |\n| Stage 4' | **checkpoint** | Stage 4' completed | Wait for user confirmation |\n| checkpoint | Stage 4.5 | User confirms | Pass revised draft to final verification |\n| Stage 4.5 | **checkpoint** | PASS (zero issues) | Wait for user confirmation |\n| Stage 4.5 | Stage 4.5 (retry) | FAIL | Fix issues, re-verify (max 3 rounds) |\n| checkpoint | Stage 5 | User confirms | Pass final accepted draft |\n\n### Special Flow Transitions\n\n| From | To | Precondition | Action |\n|------|----|-------------|--------|\n| Stage 3 (Reject) | Stage 2 | User chooses to restructure | Clear Stage 2-3 state, preserve Stage 1 materials, restart Stage 2 |\n| Stage 3 (Reject) | ABORT | User chooses to abandon | Save all produced materials, mark pipeline aborted |\n| Stage 3' (Major) | Stage 4' | User confirms | Last revision opportunity |\n| Stage 4' | Stage 4.5 | Revision complete | Go directly to final verification (no return to review) |\n| Any stage | PAUSED | User says \"pause\" or \"stop here\" | Save pipeline state |\n| PAUSED | Previous stage | User returns to continue | Restore pipeline state, display Dashboard |\n\n### Prohibited Transitions (Illegal)\n\n| From | To | Reason |\n|------|----|--------|\n| Stage 1 | Stage 3 | Cannot skip Stage 2 and 2.5 (unless mid-entry + has paper) |\n| Stage 2 | Stage 3 | **Cannot skip Stage 2.5 (integrity check is mandatory)** |\n| Stage 4 | Stage 5 | Cannot skip RE-REVIEW (revision must be re-reviewed) |\n| Stage 3' | Stage 5 | **Cannot skip Stage 4.5 (final integrity check is mandatory)** |\n| Stage 4' | Stage 3' | Cannot return to RE-REVIEW (max 1 round of RE-REVISE) |\n| Stage 5 | Stage 3 | Cannot roll back (no review after FINALIZE) |\n| completed | in_progress | Completed stages cannot restart |\n\n---\n\n## Material Dependency Matrix\n\n| Material | Produced At | Consumed At | Required/Recommended |\n|----------|-----------|-------------|---------------------|\n| RQ Brief | Stage 1 | Stage 2 (Phase 0) | Recommended |\n| Methodology Blueprint | Stage 1 | Stage 2 (Phase 0) | Recommended |\n| Bibliography | Stage 1 | Stage 2 (Phase 1) | Recommended |\n| Synthesis Report | Stage 1 | Stage 2 (Phase 3) | Recommended |\n| Paper Draft | Stage 2 | Stage 2.5 (input) | **Required** |\n| **Integrity Report (Pre)** | **Stage 2.5** | **Stage 3 (prerequisite)** | **Required** |\n| **Verified Paper Draft** | **Stage 2.5** | **Stage 3 (Phase 0)** | **Required** |\n| Review Reports (x5) | Stage 3 | Stage 4 (input) | Required |\n| Editorial Decision | Stage 3 | Stage 4 (input) | Required |\n| Revision Roadmap | Stage 3 | Stage 4 (input) | Required |\n| Revised Draft | Stage 4 | Stage 3' (Phase 0) | Required |\n| Response to Reviewers | Stage 4 | Stage 3' (input) | Recommended |\n| **Re-Review Report** | **Stage 3'** | **Stage 4' (input)** | **Required (if Major)** |\n| **Re-Revised Draft** | **Stage 4'** | **Stage 4.5 (input)** | **Required (if executed)** |\n| **Integrity Report (Final)** | **Stage 4.5** | **Stage 5 (prerequisite)** | **Required** |\n| Final Paper | Stage 5 | END (delivery) | Required |\n\n---\n\n## Exception State Handling\n\n### Timeout\n\nIf a stage shows no progress for an extended period (e.g., Socratic mode exceeds 15 rounds without convergence):\n1. state_tracker marks the stage as `stalled`\n2. orchestrator provides options:\n   - Switch mode (socratic -> full)\n   - Narrow scope\n   - Skip this stage (non-mandatory stages only)\n\n### Missing Materials\n\nIf required materials are found missing during transition:\n1. state_tracker reports the material gap\n2. orchestrator suggests returning to the stage that produces that material\n3. User can choose: backfill / skip (at own risk, but cannot skip integrity checks)\n\n### Integrity Check FAIL Loop\n\nIf Stage 2.5 or 4.5 corrections exceed 3 rounds without passing:\n1. List all unverifiable items\n2. User decides:\n   - Manually handle unverifiable items\n   - Remove unverifiable citations\n   - Continue to next stage (with \"partially unverified\" warning)\n\n### Session Interruption\n\nIf the user leaves and returns:\n1. orchestrator displays Progress Dashboard\n2. Confirm whether to continue from breakpoint\n3. Check if any outdated materials need refreshing\n\n---\n\n## Revision Loop Rules (v2.0)\n\n### Simplified Revision Cycle\n\n```\nv2.0's revision cycle is simpler and more explicit than v1.0:\n\nStage 3 (First REVIEW)\n  -> Decision: Accept -> Stage 4.5\n  -> Decision: Minor/Major -> Stage 4\n      -> Stage 4 (REVISE)\n          -> Stage 3' (RE-REVIEW, verification)\n              -> Decision: Accept/Minor -> Stage 4.5\n              -> Decision: Major -> Stage 4' (last revision)\n                  -> Stage 4.5 (go directly to final verification, no return to review)\n\nMaximum 1 round of RE-REVISE, no infinite loops.\nUnresolved issues -> Acknowledged Limitations.\n```\n\n### Differences from v1.0\n\n| v1.0 | v2.0 |\n|------|------|\n| Max 2 review-revise cycles | Fixed 2 reviews (Stage 3 + Stage 3') + max 1 RE-REVISE |\n| No integrity check | Mandatory Pre-review + Final integrity check |\n| 4 reviewers | 5 reviewers (+Devil's Advocate) |\n| Can skip any stage | Stage 2.5 and 4.5 cannot be skipped |\n| No mandatory checkpoints | Every stage requires a checkpoint |\n\nFile v1.0.0:references/plagiarism_detection_protocol.md\n\n# Plagiarism Detection Protocol — Phase D Originality Verification Protocol\n\nThis document defines the complete execution protocol for `integrity_verification_agent`'s Phase D (originality verification), including paragraph-level comparison, self-plagiarism check, AI-generated text characteristic detection, severity grading, and tool limitation disclaimers.\n\n---\n\n## Phase D Overview: Originality Verification\n\nPhase D's purpose is to perform originality screening on body text content before paper submission for review and after revision completion. Unlike Phases A-C which focus on \"whether citations and data are correct,\" Phase D focuses on \"whether the body text itself is originally written.\"\n\n**Core principle: Heuristic screening, not final determination.** This protocol uses WebSearch for publicly available literature comparison. Results are preliminary screening signals and do not equate to conclusions from professional plagiarism detection software.\n\n---\n\n## D1: Paragraph-Level Originality Check\n\n### D1.1 Characteristic Sentence Extraction\n\n```\nFor each paragraph in the paper body text:\n1. Identify the paragraph's topic and core argument\n2. Extract 1-2 \"characteristic sentences\"\n   - Priority selection: Sentences containing specific data, proper nouns, or unique arguments\n   - Avoid: Generic academic boilerplate (e.g., \"This study aims to...\")\n3. Record the characteristic sentence's paragraph location (section + paragraph number)\n```\n\n### D1.2 WebSearch Comparison\n\n```\nFor each extracted characteristic sentence:\n1. Use the characteristic sentence (or key fragment) as a WebSearch query\n   - Search term: Enclose 8-12 consecutive words in quotation marks\n   - Supplementary search: Remove quotes to check for paraphrased versions\n2. Review the top 5-10 search results\n3. Compare text similarity between original and search results\n```\n\n### D1.3 Comparison Result Grading\n\n| Grade | Code | Definition | Determination Criteria |\n|-------|------|-----------|----------------------|\n| Original | `ORIGINAL` | No similar expression found in public literature | WebSearch returns no related matches |\n| Common Knowledge | `COMMON_KNOWLEDGE` | The knowledge is a widely accepted fact in the field | Multiple sources express the same fact in different ways |\n| Paraphrase | `PARAPHRASE` | Expresses same viewpoint as a source but with clearly different wording | Semantically similar but significantly different sentence structure and word choice, with citation |\n| Close Match | `CLOSE_MATCH` | Highly similar wording to a source, with only a few words substituted | Nearly identical sentence structure, with only synonym substitutions or word order changes |\n| Verbatim | `VERBATIM` | Identical or nearly identical text to a source | 20+ consecutive identical words without quotation marks |\n\n### D1.4 Sampling Rate Requirements\n\n| Operating Mode | Minimum Sampling Rate | Description |\n|---------------|----------------------|------------|\n| Mode 1 (pre-review) | **30%** | At least 30% of all body text paragraphs checked |\n| Mode 2 (final-check) | **50%** | At least 50% of all body text paragraphs checked |\n\n**Sampling strategy**:\n- Priority check: Literature Review, Background, Discussion and other high-risk sections\n- Must cover: At least 1 paragraph sampled from each major chapter\n- Random supplement: Beyond priority paragraphs, randomly sample paragraphs to reach minimum sampling rate\n- Revised paragraphs: In Mode 2, all paragraphs newly added or substantially modified during revision must be checked 100%\n\n---\n\n## D2: Self-Plagiarism Check\n\n### D2.1 Author's Existing Publications Search\n\n```\nPrerequisite: User provides author name(s)\n\nFor each author (or primary author):\n1. WebSearch: \"author name\" + research area keywords\n2. Identify author's existing publication list (Google Scholar profile preferred)\n3. Record existing publications related to the current paper's topic\n```\n\n### D2.2 Comparison Items\n\n```\nCompare current paper with author's existing publications (focus on these areas):\n1. Methodology descriptions:\n   - Is the research design description verbatim identical to prior work?\n   - Are data collection and analysis method descriptions directly copied?\n2. Results narratives:\n   - Are textual descriptions of results reused?\n   - Are table/figure description texts identical?\n3. Theoretical framework:\n   - Are literature review paragraphs transferred wholesale?\n```\n\n### D2.3 Legitimate Self-Citation vs. Self-Plagiarism Determination Criteria\n\n| Scenario | Determination | Description |\n|----------|--------------|------------|\n| Cites prior work and restates in new language | **Legitimate self-citation** | Normal academic practice — has citation and paraphrasing |\n| Cites prior work but verbatim copies original text | **Self-plagiarism** | Even with citation, extensive verbatim copying is unacceptable |\n| Content highly similar to prior work without citing it | **Self-plagiarism** | Conceals relationship with prior work |\n| Uses prior work's data but re-analyzes | **Legitimate** | Secondary analysis is a legitimate research method — must clearly state this |\n| Methodology reuses prior work's standardized description | **Gray area** | Standardized experimental procedure descriptions allow high similarity, but citing prior work is recommended |\n\n---\n\n## D3: AI-Generated Text Characteristic Detection\n\n**Important disclaimer: This section is a checklist, not a determination tool.** AI text detection technology is not yet mature, and any judgment based on text characteristics has a high false-positive risk. The following indicators are for reference only and should not serve as the basis for final determination.\n\n### D3.1 Typical AI Writing Pattern Indicators\n\n| # | Indicator | Description | Observation Method |\n|---|-----------|------------|-------------------|\n| 1 | Excessive smoothness | Abnormally uniform sentence fluency throughout, lacking natural writing rhythm variation | Compare whether writing style across chapters is overly consistent |\n| 2 | Lack of specificity | Arguments remain at conceptual level, lacking specific numbers, cases, or personal research experience | Check for \"for example\" followed by vague content |\n| 3 | Formulaic transitions | Heavy use of \"Furthermore,\" \"Moreover,\" \"It is worth noting that\" and similar transitions | Count the variety and frequency of transition phrases |\n| 4 | Excessive parallelism | Highly symmetric paragraph structures (e.g., every paragraph follows: claim -> evidence -> summary) | Observe whether paragraph structure mechanically repeats |\n| 5 | Hedging overload | Excessive use of \"may,\" \"could,\" \"might,\" \"it is possible that\" to avoid definitive positions | Check whether author over-hedges even on their own research results |\n| 6 | Citation-argument gap | Literature is cited but the cited content is not organically integrated with the author's arguments | Remove citations — does the paragraph's argument still hold? |\n\n### D3.2 Handling Approach\n\n```\nIf the paper triggers 2 or more AI writing indicators:\n1. Flag in the verification report as \"AI writing characteristic alert\"\n2. List the specific indicators triggered and corresponding paragraphs\n3. Do NOT make a \"whether it is AI-generated\" determination\n4. Recommend the user review the flagged paragraphs and consider adjusting writing style\n```\n\n---\n\n## Severity Grading\n\n| Level | Code | Definition | Trigger Conditions |\n|-------|------|-----------|-------------------|\n| **Critical** | `CRITICAL` | Severe academic misconduct, sufficient for retraction | Verbatim plagiarism (>20 consecutive identical words without citation); fabricated citations (citing nonexistent sources to support plagiarized content) |\n| **Serious** | `SERIOUS` | Significant originality problems, requiring major revisions | Multiple close paraphrases without citing sources; extensive undisclosed self-plagiarism |\n| **Moderate** | `MODERATE` | Individual paragraphs need rewriting | Individual paragraphs inadequately paraphrased (1-2 instances of `CLOSE_MATCH`); methodology description overly similar to prior work |\n| **Minor** | `MINOR` | Does not affect academic integrity but improvement recommended | Excessive generic academic boilerplate; AI writing characteristic alerts (informational only, does not affect verdict) |\n\n### Severity-to-Verdict Mapping\n\n| Severity | Impact on Verdict |\n|---------|-----------------|\n| CRITICAL | Immediate FAIL, listed as highest priority correction item |\n| SERIOUS | FAIL, must fix and re-verify |\n| MODERATE | FAIL, must fix |\n| MINOR | Does not affect PASS/FAIL verdict, noted in report |\n\n---\n\n## Tool Limitation Disclaimer\n\nThis protocol's originality verification has the following inherent limitations that users must be aware of:\n\n| # | Limitation | Description |\n|---|-----------|------------|\n| 1 | **Not professional plagiarism detection software** | This protocol uses WebSearch for heuristic comparison, not Turnitin, iThenticate, or other professional tools — cannot calculate precise text overlap rates |\n| 2 | **Limited coverage** | Can only compare publicly searchable literature (open access, preprints, web pages) — cannot search full-text databases behind paywalls |\n| 3 | **Language limitation** | Cross-language plagiarism (e.g., plagiarism via translation) is difficult to detect |\n| 4 | **Sampling, not exhaustive** | Limited by efficiency, only 30%-50% of paragraphs are sampled — missed detection risk exists |\n| 5 | **Time sensitivity** | Search results change over time; newly published literature may not be in search scope |\n| 6 | **AI detection unreliable** | D3's AI writing indicators are heuristic alerts with high false-positive rates and should not serve as determination basis |\n\n**Recommendation**: This protocol's results serve as preliminary screening. It is recommended to use professional plagiarism detection tools (such as Turnitin / iThenticate) for complete duplicate checking before formal submission.\n\n---\n\n## Output Format Template\n\n```markdown\n## Phase D: Originality Verification Results\n\n### Verification Parameters\n- Operating mode: [Mode 1 pre-review / Mode 2 final-check]\n- Total body text paragraphs: X\n- Paragraphs sampled: Y (sampling rate: Z%)\n- Author self-plagiarism check: [Executed / Not executed (author information not provided)]\n\n### D1 Paragraph-Level Comparison Results Summary\n\n| Grade | Paragraph Count | Proportion |\n|-------|----------------|-----------|\n| ORIGINAL | X | X% |\n| COMMON_KNOWLEDGE | X | X% |\n| PARAPHRASE | X | X% |\n| CLOSE_MATCH | X | X% |\n| VERBATIM | X | X% |\n\n### D2 Self-Plagiarism Check Results\n\n| # | Current Paper Paragraph | Existing Publication | Similarity Type | Determination |\n|---|------------------------|---------------------|----------------|--------------|\n| 1 | §X.X, paragraph Y | Author (Year), Title | Methodology description similar | Legitimate self-citation / Self-plagiarism |\n\n### D3 AI Writing Characteristic Alerts\n\n| # | Indicator | Triggered Paragraph | Description |\n|---|-----------|---------------------|------------|\n| 1 | [Indicator name] | §X.X | [Specific observation] |\n\nIndicators triggered: X / 6 ([Below threshold, not flagged / Threshold reached, user review recommended])\n\n### Phase D Issue List\n\n| # | Severity | Type | Location | Issue Description | Matching Source | Recommended Action |\n|---|----------|------|----------|------------------|----------------|-------------------|\n| 1 | CRITICAL | VERBATIM | §X.X, paragraph Y | N consecutive words identical to source | [URL] | Rewrite or add quotation marks for direct quote |\n| 2 | SERIOUS | CLOSE_MATCH | §X.X, paragraph Y | Highly similar wording, only a few words substituted | [URL] | Rewrite and add citation |\n| 3 | MODERATE | Self-plagiarism | §X.X, paragraph Y | Methodology description verbatim identical to prior work | Author (Year) | Rewrite and cite prior work |\n\n### Tool Limitation Disclaimer\n\n> This originality verification uses WebSearch for heuristic comparison and is not professional plagiarism detection software (such as Turnitin / iThenticate). Coverage is limited to publicly searchable literature, with a sampling rate of [Z]%, and there is a risk of missed detection. These results serve as preliminary screening; it is recommended to use professional plagiarism detection tools for complete duplicate checking before formal submission.\n```\n\n---\n\n## Relationship with Other Phases\n\n| Phase | Focus | Relationship with Phase D |\n|-------|-------|--------------------------|\n| Phase A: Reference Verification | Whether references exist and are correct | A verifies sources, D verifies body text; if D finds VERBATIM without citation, it may also reveal A3 dangling citation issues |\n| Phase B: Citation Context Verification | Whether citations accurately reflect original text | B checks \"whether cited content is correct,\" D checks \"whether uncited content is original\" |\n| Phase C: Data Verification | Whether statistical data is correct | C and D are complementary: C verifies data, D verifies text |\n\n---\n\n## Reproducibility Requirements\n\nTo ensure the originality verification process is reproducible:\n\n1. **Standardized search strategy**: Use the same search template for each characteristic sentence\n   - Search term 1: `\"key fragment\" (8-12 words, in quotes)`\n   - Search term 2: `keyword combination (without quotes, to match paraphrases)`\n\n2. **Explicit determination criteria**: Each grade (ORIGINAL through VERBATIM) has clear determination criteria, not relying on subjective feeling\n\n3. **Complete records**: Search terms, search results, and determination rationale for each sampled paragraph are recorded in the Audit Trail\n\n4. **Timestamps**: Report includes execution time, as search results change over time\n\nFile v1.0.0:references/process_summary_protocol.md\n\n# Stage 6: Process Summary Protocol (Added in v2.4)\n\n**Trigger**: After Stage 5 (FINALIZE) completion\n**Purpose**: Document the complete human-AI collaboration history for the paper creation process, for user sharing, reporting, or reflection\n\n## Workflow\n\n```\n1. Ask user language preference:\n   \"Which language version of the process record would you like to generate first?\"\n   - Chinese (Traditional Chinese)\n   - English\n   - Both (default: generate the user's primary conversation language first)\n\n2. Review session history and compile the following:\n   - User's initial instructions (verbatim quote)\n   - Key decision points and user interventions at each stage\n   - Direction correction moments and reasons\n   - Iteration count and review result summaries\n   - Intellectual insights raised by the user (e.g., questions that spawned new chapters)\n   - Quality requirement evolution (e.g., formatting, tone adjustments)\n   - Pipeline statistics (stage count, review rounds, integrity verification count, etc.)\n\n3. Generate Markdown version (paper_creation_process.md / paper_creation_process_en.md)\n\n4. Convert to LaTeX and compile PDF:\n   - pandoc MD -> LaTeX body\n   - Package complete LaTeX document (with cover page, table of contents, headers/footers)\n   - tectonic compile PDF\n   - Chinese version requires xeCJK + Source Han Serif TC VF\n```\n\n## Required Content in Process Record\n\n| Section | Content |\n|---------|---------|\n| Paper Information | Title, final deliverables list |\n| Stage-by-Stage Process | Input/output/key decisions for each stage, with verbatim user quotes |\n| Iteration Details | Review comment summaries, revision items, re-review results |\n| Interaction Pattern Summary | User role, Claude role, intervention count, key turning points — statistics table |\n| User Key Decisions | Chronological list of every important decision made by the user |\n| Key Lessons | Reusable lessons learned from the process |\n| **Collaboration Quality Evaluation** | **Final chapter: 1-100 score + dimensional analysis + improvement suggestions** (see below) |\n\n## Collaboration Quality Evaluation (Final Chapter, Mandatory)\n\nThe final chapter of the process record is a \"Collaboration Quality Evaluation\" that honestly and constructively assesses the user's performance in the human-AI collaboration. Format follows the Claude Code CLI `/insight` feature.\n\n### Scoring Dimensions (each 1-100, weighted average for overall score)\n\n```\n+--------------------------------------------------+\n|  Collaboration Quality Score: [XX]/100            |\n+--------------------------------------------------+\n|                                                   |\n|  Direction Setting          [----------  ] XX     |\n|  Clarity, timing, scope definition                |\n|                                                   |\n|  Intellectual Contribution  [------------ ] XX    |\n|  Insight depth, original questions, concept        |\n|  challenges                                       |\n|                                                   |\n|  Quality Gatekeeping        [---------   ] XX     |\n|  Visual inspection, formatting requirements,       |\n|  quality standards                                |\n|                                                   |\n|  Iteration Discipline       [----------  ] XX     |\n|  Timely direction correction, willingness to       |\n|  re-run pipeline, refusing to settle              |\n|                                                   |\n|  Delegation Efficiency      [-------     ] XX     |\n|  When to intervene/when to let go, instruction     |\n|  precision, checkpoint efficiency                 |\n|                                                   |\n|  Meta-Learning              [------------ ] XX    |\n|  Feeding experience back to skills, requesting     |\n|  lesson recording, process improvement awareness  |\n|                                                   |\n+--------------------------------------------------+\n```\n\n### Scoring Criteria\n\n| Score Range | Meaning |\n|------------|---------|\n| 90-100 | Exceptional — User intervention significantly elevated the paper's intellectual quality beyond what AI could produce independently |\n| 75-89 | Excellent — User made correct directional decisions and effectively leveraged the pipeline's iteration capabilities |\n| 60-74 | Good — User completed necessary decisions but some opportunities were missed |\n| 40-59 | Basic — User primarily served as a \"continue\" button with little substantive intervention |\n| 1-39 | Needs Improvement — User intervention may have disrupted the workflow or lacked critical quality gatekeeping |\n\n### Required Subsections\n\n1. **Overall Score**: Total score + one-sentence evaluation\n2. **What Worked Well**: 2-4 specific behaviors, with verbatim user quotes\n3. **Missed Opportunities**: 1-3 things the user could have done but didn't\n4. **Recommendations for Next Time**: 3-5 specific, actionable improvement suggestions\n5. **Human vs AI Value-Add**: Clearly identify which aspects of the final paper quality came from user intervention (not achievable by AI independently)\n\n### Evaluation Principles\n\n- **Honesty first**: No inflation, no pleasantries. If the user only pressed \"continue,\" reflect that truthfully\n- **Evidence-based**: Every score is supported by specific behaviors or conversation records\n- **Constructive**: Every criticism must include actionable improvement suggestions\n- **Acknowledge uncertainty**: If certain dimensions cannot be evaluated (e.g., mid-entry skipped the research stage), mark as N/A\n- **Bidirectional reflection**: Also candidly point out Claude's shortcomings during the process (e.g., areas requiring multiple corrections)\n\n## AI Self-Reflection Report (Mandatory)\n\nThe second-to-last chapter of the process record is an \"AI Self-Reflection Report\" that honestly documents AI's own behavioral patterns during the pipeline. This complements the Collaboration Quality Evaluation (which assesses the user) by assessing the AI.\n\n### Tracked Metrics\n\nAll metrics below are derived from existing agent logs (`[DA-DECISION]`, `[DA-REBUTTAL]`, `[HEALTH-CHECK]`, state tracker JSON) — no additional per-stage instrumentation is required. The orchestrator aggregates these at Stage 6 by scanning the dialogue transcript:\n\n```\n+--------------------------------------------------+\n|  AI Self-Reflection Report                        |\n+--------------------------------------------------+\n|                                                   |\n|  DA Concession Rate           X/Y (Z%)           |\n|  (concessions / total rebuttals received)         |\n|                                                   |\n|  DA Consecutive Concessions   [list if any]       |\n|  (violations of no-consecutive rule)              |\n|                                                   |\n|  Checkpoints Skipped          X/Y                 |\n|  (SLIM or user-skipped / total checkpoints)       |\n|                                                   |\n|  User Overrides               X                   |\n|  (times user overruled AI recommendation)         |\n|                                                   |\n|  Dialogue Health Alerts       X                   |\n|  (health check interventions triggered)           |\n|  - Persistent Agreement:      X                   |\n|  - Conflict Avoidance:        X                   |\n|  - Premature Convergence:     X                   |\n|                                                   |\n|  Intent Mode Transitions      X                   |\n|  (exploratory ↔ goal-oriented switches)           |\n|                                                   |\n|  Cross-Model Disagreements    X (if enabled)      |\n|  (integrity + DA combined)                        |\n|                                                   |\n+--------------------------------------------------+\n```\n\n### Required Subsections\n\n1. **Behavioral Summary**: One paragraph describing the overall AI behavioral pattern during this pipeline run\n2. **Sycophancy Risk Assessment**: Screening thresholds based on concession rate and health alerts — LOW (concession <50%, 0 health alerts) / MEDIUM (50-65% or 1-2 alerts) / HIGH (>65% or 3+ alerts). These are screening thresholds, not diagnostic criteria — a MEDIUM rating means the metrics warrant human review, not that sycophancy occurred (a high concession rate may reflect genuinely strong rebuttals). If HIGH, include a warning: \"AI may have been too accommodating in this run. Human review of DA findings and integrity results is strongly recommended.\"\n3. **Frame-Lock Incidents**: List any `[CROSS-MODEL-FINDING]` that the primary DA missed (if cross-model was enabled), or any frame-lock detections triggered during checkpoints. If none, state \"No frame-lock incidents detected — note this could mean either good coverage or undetected frame-lock.\"\n4. **Convergence Pattern**: In Socratic dialogue stages, was intent correctly detected? Did the mentor try to converge prematurely? Report mode transitions and any premature-convergence health alerts.\n5. **What AI Got Wrong**: Candid list of AI errors or shortcomings during the run — corrections needed, checkpoint failures, integrity issues found. This is not a failure report; it is evidence that quality gates are working.\n6. **Failure Mode Audit Log** (v3.2): For each of the 7 AI research failure modes from the Stage 2.5 / 4.5 checklist (see `references/ai_research_failure_modes.md`), report (a) final status at 4.5 — `CLEAR` / `OVERRIDDEN`, (b) history — was it ever `SUSPECTED` during the pipeline? At which stage? How was it resolved? (c) if `OVERRIDDEN`, the user's recorded reasoning. This makes the failure-mode defences part of the permanent process record. Modes with no history can be listed as `CLEAR (no flags)` in one line; expand only on modes that were flagged.\n\n### Output Length Guidance\n\nFor dimensions with no findings, state the null result in one sentence. Expand only when issues are detected. The real risk is generating verbose \"everything is fine\" paragraphs for empty subsections — resist this.\n\n### Principles\n\n- **Self-honesty**: AI must not minimize its own shortcomings. If the DA conceded too easily, say so.\n- **Not self-flagellation**: The purpose is transparency, not performative humility. Report facts with interpretation.\n- **Actionable**: Every finding should suggest what could be done differently next time (e.g., \"Consider enabling cross-model verification for the next run\" or \"The user might want to push back harder on DA concessions\")\n- **The irony is noted**: This self-reflection is itself produced by the same AI that may have been sycophantic during the pipeline. The user should read it with that awareness. This caveat must be stated in the report.\n\n## Output Specifications\n\n- **Filename**: `paper_creation_process.md` (Chinese) / `paper_creation_process_en.md` (English)\n- **PDF**: `paper_creation_process_zh.pdf` / `paper_creation_process_en.pdf`\n- **LaTeX template**: `article` class, 12pt, A4, Times New Roman + Source Han Serif TC VF\n- **Includes table of contents**: `\\tableofcontents`\n- **Header**: left = document title (italic), right = date\n- **Compilation**: tectonic (same toolchain as Stage 5)\n\nFile v1.0.0:references/progress_dashboard_template.md\n\n# Progress Dashboard\n\nUsers can say \"status\" or \"pipeline status\" at any time to view:\n\n```\n+=============================================+\n|   Academic Pipeline Status                   |\n+=============================================+\n| Topic: Impact of AI on Higher Education     |\n|        Quality Assurance                    |\n+---------------------------------------------+\n\n  Stage 1   RESEARCH          [v] Completed\n  Stage 2   WRITE             [v] Completed\n  Stage 2.5 INTEGRITY         [v] PASS (62/62 refs verified)\n  Stage 3   REVIEW (1st)      [v] Major Revision (5 items)\n  Stage 4   REVISE            [v] Completed (5/5 addressed)\n  Stage 3'  RE-REVIEW (2nd)   [v] Accept\n  Stage 4'  RE-REVISE         [-] Skipped (Accept)\n  Stage 4.5 FINAL INTEGRITY   [..] In Progress\n  Stage 5   FINALIZE          [ ] Pending\n  Stage 6   PROCESS SUMMARY   [ ] Pending\n\n+---------------------------------------------+\n| Integrity Verification:                     |\n|   Pre-review:  PASS (0 issues)              |\n|   Final:       In progress...               |\n+---------------------------------------------+\n| Review History:                             |\n|   Round 1: Major Revision (5 required)      |\n|   Round 2: Accept                           |\n+=============================================+\n```\n\nSee `templates/pipeline_status_template.md` for the output template.","readmeExcerpt":"Skill: academic-pipeline-v1 Owner: eric-promax Summary: Orchestrator for the full academic research pipeline: literature search -> research -> write -> integrity check -> review -> revise -> re-review -> re-revise... Tags: latest:1.0.0 Version history: v1.0.0 | 2026-05-14T02:46:10.557Z | user Academic Pipeline v1.0.0 — Initial Release - Introduces a full academic research workflow orchestrator covering 12 stages from","codeSnippets":[],"executableExamples":[{"language":"text","snippet":"I want to write a research paper on the impact of AI on higher education quality assurance"},{"language":"text","snippet":"I already have a paper, help me review it"},{"language":"text","snippet":"I received reviewer comments, help me revise"},{"language":"text","snippet":"━━━ Stage [X] [Name] Complete ━━━\n\nMetrics:\n- Word count: [N] (target: [T] +/-10%)    [OK/OVER/UNDER]\n- References: [N] (min: [M])              [OK/LOW]\n- Coverage: [N]/[T] sections drafted       [COMPLETE/PARTIAL]\n- Quality indicators: [score if available]\n\nDeliverables:\n- [Material 1]\n- [Material 2]\n\nFlagged: [any issues detected, or \"None\"]\n\nReady to proceed to Stage [Y]? You can also:\n1. View progress (say \"status\")\n2. Adjust settings\n3. Pause pipeline\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"},{"language":"text","snippet":"pipeline_orchestrator_agent analyzes the user's input:\n\n1. What materials does the user have?\n   - No materials           --> Stage 2 (RESEARCH)\n   - Has research data      --> Stage 3 (WRITE)\n   - Has paper draft        --> Stage 4 (INTEGRITY)\n   - Has verified paper     --> Stage 5 (REVIEW)\n   - Has review comments    --> Stage 7 (REVISE)\n   - Has revised draft      --> Stage 6 (RE-REVIEW)\n   - Has final draft for formatting --> Stage 11 (FINALIZE)\n\n2. What is the user's goal?\n   - Full workflow (research to publication)\n   - Partial workflow (only certain stages needed)\n\n3. Determine entry point, confirm with user"},{"language":"text","snippet":"Based on entry point and user preferences, recommend modes for each stage:\n\nUser type determination:\n- Novice / wants guidance --> socratic (Stage 2) + plan (Stage 3) + guided (Stage 5)\n- Experienced / wants direct output --> full (Stage 2) + full (Stage 3) + full (Stage 5)\n- Time-limited --> quick (Stage 2) + full (Stage 3) + quick (Stage 5)\n\nExplain the differences between modes when recommending, letting the user choose"}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: academic-pipeline\ndescription: \"Orchestrator for the full academic research pipeline: literature search -> research -> write -> integrity check -> review -> revise -> re-review -> re-revise -> final integrity check -> humanize -> finalize. Coordinates academic-search, deep-research, academic-paper, academic-paper-reviewer, and humanizer into a seamless 12-stage workflow with mandatory integrity verification, two-stage peer review, de-AI processing, and reproducible quality gates. Triggers on: academic pipeline, research to paper, full paper workflow, paper pipeline, end-to-end paper, research-to-publication, complete paper workflow.\"\nmetadata:\n  version: \"3.5\"\n  last_updated: \"2026-05-13\"\n  depends_on: \"ima-skills, academic-search, deep-research, academic-paper, academic-paper-reviewer, humanizer, humanizer-zh\"\n  status: active\n  related_skills:\n    - ima-skills\n    - academic-search\n    - deep-research\n    - academic-paper\n    - academic-paper-reviewer\n    - humanizer\n    - humanizer-zh\n---\n\n# Academic Pipeline v3.5 — Full Academic Research Workflow Orchestrator\n\nA lightweight orchestrator that manages the complete academic pipeline from research exploration to final manuscript. It does not perform substantive work — it only detects stages, recommends modes, dispatches skills, manages transitions, and tracks state.\n\n**v2.0 Core Improvements**:\n1. **Mandatory user confirmation checkpoints** — Each stage completion requires user confirmation before proceeding to the next step\n2. **Academic integrity verification** — After paper completion and before review submission, 100% reference and data verification must pass\n3. **Two-stage review** — First full review + post-revision focused verification review\n4. **Final integrity check** — After revision completion, re-verify all citations and data are 100% correct\n5. **Reproducible** — Standardized workflow producing consistent quality assurance each time\n6. **Process documentation** — After pipeline completion, automatically generates a \"Paper Creation Process Record\" PDF documenting the human-AI collaboration history\n\n## Quick Start\n\n**Full workflow (from scratch):**\n```\nI want to write a research paper on the impact of AI on higher education quality assurance\n```\n--> academic-pipeline launches, starting from Stage 2 (RESEARCH)\n\n**Mid-entry (existing paper):**\n```\nI already have a paper, help me review it\n```\n--> academic-pipeline detects mid-entry, starting from Stage 4 (INTEGRITY)\n\n**Revision mode (received reviewer feedback):**\n```\nI received reviewer comments, help me revise\n```\n--> academic-pipeline detects, starting from Stage 7 (REVISE)\n\n**Execution flow:**\n1. Detect the user's current stage and available materials\n2. Recommend the optimal mode for each stage\n3. Dispatch the corresponding skill for each stage\n4. **After each stage completion, proactively prompt and wait for user confirmation**\n5. Track progress throughout; Pipeline Status Dashboard available at any time\n\n---\n\n## Trigger"},{"path":"_meta.json","content":"{\n  \"ownerId\": \"kn79h8an2b3e37qcycc96jzpjs86qv7g\",\n  \"slug\": \"academic-pipeline\",\n  \"version\": \"1.0.0\",\n  \"publishedAt\": 1778726770557\n}"},{"path":"references/ai_research_failure_modes.md","content":"# AI Research Failure Mode Checklist\n\n**Status**: v3.2\n**Parent skill**: `academic-pipeline`\n**Used at**: Stage 2.5 INTEGRITY (blocking), Stage 4.5 FINAL INTEGRITY (blocking), Stage 6 PROCESS SUMMARY (reporting only)\n**Source**: Lu et al. (2026). Towards end-to-end automation of AI research. *Nature* 651, 914-919. doi:10.1038/s41586-026-10265-5 — Limitations section, Figure 2 (examples of failures in The AI Scientist's own accepted paper), Supplementary Information A.2.9 (debugging traces).\n\n---\n\n## Why this checklist exists\n\nLu et al. built the first autonomous AI research system to pass blind peer review (ICLR 2025 workshop). Their Limitations section enumerates the specific failure modes they observed — and most of them apply equally to human-in-the-loop AI research workflows like ARS.\n\nThese failures are dangerous because **they look like competent work**. A paper containing a hallucinated experimental result reads the same as a paper containing a real one. A shortcut-relying result reads the same as a genuine generalization. A methodology section describing experiments that were never actually run reads the same as a faithful account. The existing integrity verification catches citation hallucinations but is weak on the other failure modes.\n\nThe checklist exists to make these failures legible: **at Stage 2.5 and Stage 4.5, the integrity reviewer must explicitly rule out each of the 7 modes, or flag which are suspected and block the pipeline until the user acknowledges.**\n\nThis also extends the existing 5-type citation hallucination taxonomy (in `academic-paper-reviewer` references) into a broader 7-type AI research hallucination taxonomy. Citation hallucinations become mode 2 below.\n\n---\n\n## The 7 failure modes\n\n### Mode 1: Implementation bug passing AI self-review\n\n**What it is**: The analysis or experiment code has a bug (off-by-one, wrong variable, silent division-by-zero, type coercion, wrong flag) that produces numerically plausible but scientifically wrong results. The AI runs the code, looks at the output, sees nothing \"obviously\" wrong, and incorporates the result into the paper.\n\n**Lu 2026 example**: Supplementary A.2.9 traces show The AI Scientist repeatedly accepting experimental runs that had silent crashes or numerical instabilities because the top-level metric \"looked reasonable\". Figure 2 shows an ICLR reviewer catching one such issue in the accepted paper — the paper's main analysis depended on a setup that the code did not actually implement.\n\n**Detection questions at Stage 2.5**:\n- For every numerical result in the draft: does the user have a saved log, notebook, or script run that produced this number? If yes, was the exit code 0 and were there zero warnings? If no log is saved, flag.\n- Are any effect sizes suspiciously round (exactly 0.5, exactly 2x baseline, exactly zero variance across runs)? Suspiciously round numbers are a common signal of a constant leaking through a broken pipeline.\n- Do error bars / confidence inte"},{"path":"references/changelog.md","content":"# Changelog\n\n| Version | Date | Changes |\n|---------|------|---------|\n| 2.7 | 2026-03-27 | **Style Profile in Material Passport**: Pipeline orchestrator now carries optional Style Profile (Schema 10 in `shared/handoff_schemas.md`) through all stages. Produced by academic-paper intake Step 10 when user provides past writing samples. Consumed by draft_writer (Stage 2) and report_compiler (Stage 1) as soft writing voice guide. Does not affect integrity verification or review stages. Coordinates with deep-research v2.4 and academic-paper v2.5 |\n| 2.6 | 2026-03-08 | **Handoff Data Schema**: Enhanced `shared/handoff_schemas.md` with 9 comprehensive schemas (RQ Brief, Bibliography, Synthesis, Paper Draft, Integrity Report, Review Report, Revision Roadmap, Response to Reviewers, Material Passport) with full field definitions, type constraints, and validation rules; orchestrator validates output against schemas before each transition. **Adaptive Checkpoint System**: Replaced static checkpoint template with 3-tier system (FULL/SLIM/MANDATORY) based on stage criticality and user engagement; FULL checkpoints include decision dashboard with metrics; SLIM auto-continues for experienced users; MANDATORY cannot be bypassed at integrity/review/finalization boundaries; awareness guard after 4+ auto-continues. **Mode Advisor**: New `references/mode_advisor.md` with unified cross-skill decision tree, common misconceptions table, user archetype recommendations, decision flowchart, and anti-patterns guide. **Team Collaboration Protocol**: New `references/team_collaboration_protocol.md` with 5 role definitions, per-transition handoff procedures, git branching/tagging strategy, conflict resolution matrix, and communication templates; state tracker extended with `assigned_to`, `approval_gate`, `team_notes` per stage and `schema_validation_log`. **Phase E Claim Verification**: New `references/claim_verification_protocol.md` with E1 claim extraction, E2 source tracing, E3 cross-referencing; verdict taxonomy (VERIFIED / MINOR_DISTORTION / MAJOR_DISTORTION / UNVERIFIABLE / UNVERIFIABLE_ACCESS); severity mapping (MAJOR_DISTORTION -> SERIOUS, UNVERIFIABLE -> SERIOUS, MINOR_DISTORTION -> MINOR, UNVERIFIABLE_ACCESS -> MEDIUM); integrated into integrity_verification_agent Mode 1 (30% spot-check) and Mode 2 (100%); pass/fail criteria updated to include Phase E verdicts. **Mid-Entry Material Passport Check**: Pipeline orchestrator now validates Material Passport on mid-entry; decision tree checks verification_status, freshness (< 24 hours), and content modification (version_label comparison); offers skip/spot-check/full re-verify options for Stage 2.5 when passport is valid; passport freshness validation rules added to `shared/handoff_schemas.md` |\n| 2.5 | 2026-03-08 | External Review Protocol: structured intake of real journal reviewer feedback (text/PDF/DOCX); 4-step workflow (parse -> strategic coaching -> revise + Response to Reviewers -> completeness check); differentiated be"},{"path":"references/claim_verification_protocol.md","content":"# Claim Verification Protocol (Phase E)\n\n## Purpose\nVerifies that quantitative and factual claims in the paper are accurately supported by their cited sources. Phase A-D verify that references exist and are original; Phase E verifies that claims derived from those references are truthful.\n\n## Scope\n- All numerical claims (percentages, counts, effect sizes, p-values)\n- All categorical assertions (\"X is the largest...\", \"Y was the first to...\")\n- All trend claims (\"increasing\", \"declining\", \"stable\")\n- All causal claims (\"X causes Y\", \"X leads to Y\")\n\n## E1: Claim Extraction\n- Scan the paper for all quantitative/factual claims\n- For each claim, record: claim text, cited source(s), paper section, page/line\n- Expected output: Claim Registry table\n\n## E2: Source Tracing\n- For each claim, locate the specific passage in the cited source that supports it\n- Use WebSearch + DOI lookup to find the original source\n- If source is behind paywall, note as UNVERIFIABLE_ACCESS\n\n## E3: Cross-Referencing\n- Compare claim text vs source text\n- Check: exact numbers, date ranges, population descriptions, methodology descriptions\n- Flag any discrepancies\n\n## Verdict Taxonomy\n\n| Verdict | Definition | Severity | Example |\n|---------|-----------|----------|---------|\n| VERIFIED | Claim matches source exactly or within rounding tolerance | None | Paper: \"15.2%\"; Source: \"15.2%\" |\n| MINOR_DISTORTION | Claim paraphrases source but meaning is preserved | MINOR | Paper: \"about 15%\"; Source: \"15.2%\" |\n| MAJOR_DISTORTION | Claim oversimplifies, exaggerates, or misrepresents source | SERIOUS | Paper: \"declined sharply\"; Source: \"declined by 2.1%\" |\n| UNVERIFIABLE | Source doesn't contain the claimed information | SERIOUS | Paper cites Smith (2020) for a claim, but Smith (2020) doesn't discuss this topic |\n| UNVERIFIABLE_ACCESS | Source exists but full text not accessible for verification | MEDIUM | Paywalled journal article |\n\n## Sampling Strategy\n- Mode 1 (pre-review): 30% random sample of claims (minimum 10 claims)\n- Mode 2 (final-check): 100% of claims\n\n## Output Format\n\n### Claim Verification Report\n| # | Claim | Source | Section | Verdict | Detail |\n|---|-------|-------|---------|---------|--------|\n| 1 | [claim text] | [source] | [section] | VERIFIED | Exact match |\n| 2 | [claim text] | [source] | [section] | MAJOR_DISTORTION | Paper says X, source says Y |\n\n### Summary\n- Total claims checked: [N]\n- VERIFIED: [N]\n- MINOR_DISTORTION: [N]\n- MAJOR_DISTORTION: [N] (must be 0 for PASS)\n- UNVERIFIABLE: [N] (must be 0 for PASS)\n- UNVERIFIABLE_ACCESS: [N] (noted but does not block PASS)\n\n## Pass/Fail Criteria\n- PASS: Zero MAJOR_DISTORTION + Zero UNVERIFIABLE\n- FAIL: Any MAJOR_DISTORTION or UNVERIFIABLE\n- PASS_WITH_NOTES: Only MINOR_DISTORTION and/or UNVERIFIABLE_ACCESS"}],"languages":[],"docsSourceLabel":"CLAWHUB","editorialOverview":"Orchestrator for the full academic research pipeline: literature search -> research -> write -> integrity check -> review -> revise -> re-review -> re-revise... Skill: academic-pipeline-v1 Owner: eric-promax Summary: Orchestrator for the full academic research pipeline: literature search -> research -> write -> integrity check -> review -> revise -> re-review -> re-revise... Tags: latest:1.0.0 Version history: v1.0.0 | 2026-05-14T02:46:10.557Z | user Academic Pipeline v1.0.0 — Initial Release - Introduces a full academic research workflow orchestrator covering 12 stages from","editorialQuality":{"score":100,"threshold":65,"status":"ready","wordCount":1948,"uniquenessScore":46,"reasons":[]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-11T15:53:46.384Z","emptyReason":"No screenshots, media assets, or demo links are available."},"primaryImageUrl":null,"mediaAssetCount":0,"assets":[],"demoUrl":null},"ownerResources":{"evidence":{"source":"unclaimed","verified":false,"confidence":"low","updatedAt":"2026-10-11T15:53:46.384Z","emptyReason":"This page has not been claimed by the agent owner."},"hasCustomPage":false,"customPageUpdatedAt":null,"customLinks":[],"structuredLinks":{"docsUrl":null,"demoUrl":null,"supportUrl":null,"pricingUrl":null,"statusUrl":null},"customPage":null},"relatedAgents":{"evidence":{"source":"protocol-neighbors","verified":false,"confidence":"medium","updatedAt":"2026-10-11T20:59:20.177Z","emptyReason":null},"items":[{"id":"8ebccd8e-3863-4187-8355-c3f14e1f9edf","entityType":"agent","canonicalPath":"/agent/iofficeai-aionui","slug":"iofficeai-aionui","name":"AionUi","description":"Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!","url":"https://github.com/iOfficeAI/AionUi","homepage":"https://www.aionui.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-10-09T19:11:12.944Z","createdAt":"2026-02-25T03:38:16.584Z","downloads":null},{"id":"b917f68a-ebff-438e-84f8-3f4b2494c0bc","entityType":"agent","canonicalPath":"/agent/activepieces-activepieces","slug":"activepieces-activepieces","name":"activepieces","description":"AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents","url":"https://github.com/activepieces/activepieces","homepage":"https://www.activepieces.com","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-15T02:22:12.426Z","createdAt":"2026-02-25T03:38:12.412Z","downloads":null},{"id":"5cb26759-3a39-483f-94cf-276a98c13bb8","entityType":"agent","canonicalPath":"/agent/cherryhq-cherry-studio","slug":"cherryhq-cherry-studio","name":"cherry-studio","description":"AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs","url":"https://github.com/CherryHQ/cherry-studio","homepage":"https://cherry-ai.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-11T14:38:40.986Z","createdAt":"2026-02-25T03:38:19.379Z","downloads":null},{"id":"6f6582d0-5d76-4f0f-b81d-86520247950b","entityType":"agent","canonicalPath":"/agent/copilotkit-copilotkit","slug":"copilotkit-copilotkit","name":"CopilotKit","description":"The Frontend for Agents & Generative UI. React + Angular","url":"https://github.com/CopilotKit/CopilotKit","homepage":"https://docs.copilotkit.ai","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-03-25T09:50:57.846Z","createdAt":"2026-02-25T03:39:14.617Z","downloads":null}],"links":{"hub":"/agent","source":"/agent/source/clawhub","protocols":[{"label":"OpenClaw","href":"/agent/protocol/openclew"}]}}}