Academic Pipeline v2.6 — Full Academic Research Workflow Orchestrator

A 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.

v2.0 Core Improvements:

1. Mandatory user confirmation checkpoints — Each stage completion requires user confirmation before proceeding to the next step
2. Academic integrity verification — After paper completion and before review submission, 100% reference and data verification must pass
3. Two-stage review — First full review + post-revision focused verification review
4. Final integrity check — After revision completion, re-verify all citations and data are 100% correct
5. Reproducible — Standardized workflow producing consistent quality assurance each time
6. Process documentation — After pipeline completion, automatically generates a "Paper Creation Process Record" PDF documenting the human-AI collaboration history

Quick Start

Full workflow (from scratch):

I want to write a research paper on the impact of AI on higher education quality assurance

--> academic-pipeline launches, starting from Stage 1 (RESEARCH)

Mid-entry (existing paper):

I already have a paper, help me review it

--> academic-pipeline detects mid-entry, starting from Stage 2.5 (INTEGRITY)

Revision mode (received reviewer feedback):

I received reviewer comments, help me revise

--> academic-pipeline detects, starting from Stage 4 (REVISE)

Execution flow:

1. Detect the user's current stage and available materials
2. Recommend the optimal mode for each stage
3. Dispatch the corresponding skill for each stage
4. After each stage completion, proactively prompt and wait for user confirmation
5. Track progress throughout; Pipeline Status Dashboard available at any time

Trigger Conditions

Trigger Keywords

English: academic pipeline, research to paper, full paper workflow, paper pipeline, end-to-end paper, research-to-publication, complete paper workflow

Non-Trigger Scenarios

Trigger Exclusions

• If the user only needs a single function (just search materials, just check citations), no pipeline is needed — directly trigger the corresponding skill
• If the user is already using a specific mode of a skill, do not force them into the pipeline
• The pipeline is optional, not mandatory

Pipeline Stages (10 Stages)

Pipeline State Machine

1. Stage 1 RESEARCH -> user confirmation -> Stage 2
2. Stage 2 WRITE -> user confirmation -> Stage 2.5
3. Stage 2.5 INTEGRITY -> PASS -> Stage 3 (FAIL -> fix and re-verify, max 3 rounds)
4. Stage 3 REVIEW -> Accept -> Stage 4.5 / Minor|Major -> Stage 4 / Reject -> Stage 2 or end
5. Stage 4 REVISE -> user confirmation -> Stage 3'
6. Stage 3' RE-REVIEW -> Accept|Minor -> Stage 4.5 / Major -> Stage 4'
7. Stage 4' RE-REVISE -> user confirmation -> Stage 4.5 (no return to review)
8. Stage 4.5 FINAL INTEGRITY -> PASS (zero issues) -> Stage 5 (FAIL -> fix and re-verify)
9. Stage 5 FINALIZE -> MD + DOCX -> ask about LaTeX -> confirm -> PDF -> Stage 6
10. Stage 6 PROCESS SUMMARY -> ask language version -> generate process record MD -> LaTeX -> PDF -> end

See references/pipeline_state_machine.md for complete state transition definitions.

Adaptive Checkpoint System

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.

Checkpoint Types

Decision Dashboard (shown at FULL checkpoints)

━━━ Stage [X] [Name] Complete ━━━

Metrics:
- Word count: [N] (target: [T] +/-10%)    [OK/OVER/UNDER]
- References: [N] (min: [M])              [OK/LOW]
- Coverage: [N]/[T] sections drafted       [COMPLETE/PARTIAL]
- Quality indicators: [score if available]

Deliverables:
- [Material 1]
- [Material 2]

Flagged: [any issues detected, or "None"]

Ready to proceed to Stage [Y]? You can also:
1. View progress (say "status")
2. Adjust settings
3. Pause pipeline
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Adaptive Rules

1. First checkpoint: always FULL
2. After 2+ consecutive "continue" without review: prompt user awareness ("You've auto-continued [N] times. Want to review progress?")
3. Integrity boundaries (Stage 2.5, 4.5): always MANDATORY
4. Review decisions (Stage 3, 3'): always MANDATORY
5. Before finalization (Stage 5): always MANDATORY
6. All other stages: start FULL, downgrade to SLIM if user says "just continue"

Checkpoint Rules

1. Cannot auto-skip MANDATORY checkpoints: Even if the previous stage result is perfect, explicit user input is required at MANDATORY checkpoints
2. User can adjust: At FULL and MANDATORY checkpoints, users can modify the mode or settings for the next step
3. Pause-friendly: Users can pause at any checkpoint and resume later
4. 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
5. Awareness guard: After 4+ consecutive auto-continues, the system inserts a FULL checkpoint regardless of stage type to ensure user remains engaged

Agent Team (3 Agents)

Orchestrator Workflow

Step 1: INTAKE & DETECTION

pipeline_orchestrator_agent analyzes the user's input:

1. What materials does the user have?
   - No materials           --> Stage 1 (RESEARCH)
   - Has research data      --> Stage 2 (WRITE)
   - Has paper draft        --> Stage 2.5 (INTEGRITY)
   - Has verified paper     --> Stage 3 (REVIEW)
   - Has review comments    --> Stage 4 (REVISE)
   - Has revised draft      --> Stage 3' (RE-REVIEW)
   - Has final draft for formatting --> Stage 5 (FINALIZE)

2. What is the user's goal?
   - Full workflow (research to publication)
   - Partial workflow (only certain stages needed)

3. Determine entry point, confirm with user

Step 2: MODE RECOMMENDATION

Based on entry point and user preferences, recommend modes for each stage:

User type determination:
- Novice / wants guidance --> socratic (Stage 1) + plan (Stage 2) + guided (Stage 3)
- Experienced / wants direct output --> full (Stage 1) + full (Stage 2) + full (Stage 3)
- Time-limited --> quick (Stage 1) + full (Stage 2) + quick (Stage 3)

Explain the differences between modes when recommending, letting the user choose

Step 3: STAGE EXECUTION

Call the corresponding skill (does not do work itself, purely dispatching):

1. Inform the user which Stage is about to begin
2. Load the corresponding skill's SKILL.md
3. Launch the skill with the recommended mode
4. Monitor stage completion status

After completion:
1. Compile deliverables list
2. Update pipeline state (call state_tracker_agent)
3. [MANDATORY] Proactively prompt checkpoint, wait for user confirmation

Step 4: TRANSITION

After user confirmation:

1. Pass the previous stage's deliverables as input to the next stage
2. Trigger handoff protocol (defined in each skill's SKILL.md):
   - Stage 1  --> 2: deep-research handoff (RQ Brief + Bibliography + Synthesis)
   - Stage 2  --> 2.5: Pass complete paper to integrity_verification_agent
   - Stage 2.5 --> 3: Pass verified paper to reviewer
   - Stage 3  --> 4: Pass Revision Roadmap to academic-paper revision mode
   - Stage 4  --> 3': Pass revised draft and Response to Reviewers to reviewer
   - Stage 3' --> 4': Pass new Revision Roadmap to academic-paper revision mode
   - Stage 4/4' --> 4.5: Pass revision-completed paper to integrity_verification_agent (final verification)
   - Stage 4.5 --> 5: Pass verified final draft to format-convert mode
3. Begin next stage

Integrity Review Protocol (Added in v2.0)

Stage 2.5: First Integrity Check (Pre-Review Integrity)

Trigger: After Stage 2 (WRITE) completion, before Stage 3 (REVIEW) Purpose: Ensure all references and data are not fabricated or erroneous before submission for review

Execution steps:
1. integrity_verification_agent executes Mode 1 (initial verification) on the paper
2. Verification scope:
   - Phase A: 100% reference existence + bibliographic accuracy + ghost citations
   - Phase B: >= 30% citation context spot-check
   - Phase C: 100% statistical data verification
   - Phase D: >= 30% originality spot-check + self-plagiarism check
   - Phase E: 30% claim verification spot-check (minimum 10 claims)
3. Result handling:
   - PASS -> checkpoint -> Stage 3
   - FAIL -> produce correction list -> fix item by item -> re-verify corrected items
   - PASS after corrections -> checkpoint -> Stage 3
   - Still FAIL after 3 rounds -> notify user, list unverifiable items

Stage 4.5: Final Integrity Check (Post-Revision Final Check)

Trigger: After Stage 4' (RE-REVISE) or Stage 3' (RE-REVIEW, Accept) completion, before Stage 5 (FINALIZE) Purpose: Confirm the revised paper is 100% correct and ready for publication

Execution steps:
1. integrity_verification_agent executes Mode 2 (final verification) on the revised draft
2. Verification scope:
   - Phase A: 100% reference verification (including those added during revision)
   - Phase B: 100% citation context verification (not spot-check, full check)
   - Phase C: 100% statistical data verification
   - Phase D: >= 50% originality spot-check (100% for newly added/modified paragraphs)
   - Phase E: 100% claim verification (zero MAJOR_DISTORTION + zero UNVERIFIABLE required)
3. Special check: Compare with Stage 2.5 results to confirm all previous issues are resolved
4. Result handling:
   - PASS (zero issues) -> checkpoint -> Stage 5
   - FAIL -> fix -> re-verify -> PASS -> Stage 5
5. **Must PASS with zero issues to proceed to Stage 5**

Two-Stage Review Protocol (Added in v2.0)

Stage 3: First Review (Full Review)

• Input: Paper that passed integrity check
• Review team: EIC + R1 (methodology) + R2 (domain) + R3 (interdisciplinary) + Devil's Advocate
• Output: 5 review reports + Editorial Decision + Revision Roadmap + Socratic Revision Coaching
• Decision branches: Accept -> Stage 4.5 / Minor|Major -> Revision Coaching -> Stage 4 / Reject -> Stage 2 or end

See academic-paper-reviewer/SKILL.md for review process details.

Stage 3 -> 4 Transition: Revision Coaching

EIC uses Socratic dialogue to guide the user in understanding review comments and planning revision strategy (max 8 rounds). User can say "just fix it for me" to skip.

Stage 3': Second Review (Verification Review)

• Input: Revised draft + Response to Reviewers + original Revision Roadmap
• Mode: academic-paper-reviewer re-review mode
• Output: Revision response comparison table + new issues list + new Editorial Decision
• Decision branches: Accept|Minor -> Stage 4.5 / Major -> Residual Coaching -> Stage 4'

See academic-paper-reviewer/SKILL.md Re-Review Mode for verification review process.

Stage 3' -> 4' Transition: Residual Coaching

EIC guides the user in understanding residual issues and making trade-offs (max 5 rounds). User can say "just fix it" to skip.

Mid-Entry Protocol

Users can enter from any stage. The orchestrator will:

1. Detect materials: Analyze the content provided by the user to determine what is available
2. Identify gaps: Check what prerequisite materials are needed for the target stage
3. Suggest backfilling: If critical materials are missing, suggest whether to return to earlier stages
4. Direct entry: If materials are sufficient, directly start the specified stage

Important: mid-entry cannot skip Stage 2.5

• If the user brings a paper and enters directly, go through Stage 2.5 (INTEGRITY) first before Stage 3 (REVIEW)
• Only exception: User can provide a previous integrity verification report and content has not been modified

External Review Protocol (Added in v2.5)

Scenario: The user submitted to a journal and received feedback from real human reviewers, bringing those comments into the pipeline.

Trigger: User says "I received reviewer comments," "reviewer comments," "revise and resubmit," etc.

Differences from Internal Review

Step 1: Intake and Structuring

1. Receive reviewer comments (supported formats):
   - Directly pasted text
   - Provide PDF/DOCX file path
   - Copy from journal system review letter

2. Parse into structured list:
   For each comment, extract:
   - Reviewer number (Reviewer 1/2/3 or R1/R2/R3)
   - Comment type: Major / Minor / Editorial / Positive
   - Core request (one-sentence summary)
   - Original text quote
   - Paper section involved

3. Produce External Review Summary:
   +----------------------------------------+
   | External Review Summary                |
   +----------------------------------------+
   | Journal: [journal name]                |
   | Decision: [R&R / Major / Minor]        |
   | Reviewers: [N]                         |
   | Total comments: [N]                    |
   |   Major: [n]  Minor: [n]  Editorial: [n]|
   +----------------------------------------+

4. Confirm parsing results with user:
   "I organized the reviewer comments into [N] items. Here is the summary — please confirm nothing was missed or misinterpreted."

Step 2: Strategic Revision Coaching (External Revision Coaching)

Unlike the Socratic coaching for internal review, external review coaching focuses more on strategic judgment:

For each Major comment, guide the user to think through:

1. Understanding layer
   "What is this reviewer's core concern? Is it about methodology, theory, or presentation?"

2. Judgment layer
   "Do you agree with this criticism?"
   - Agree -> "How do you plan to revise?"
   - Partially agree -> "Which parts do you agree with and which not? What is your basis for disagreement?"
   - Disagree -> "What is your rebuttal argument? Can you support it with literature or data?"

3. Strategy layer
   "How will you phrase this in the response letter?"
   - Accept revision: Show specifically what was changed and where
   - Partially accept: Explain the accepted parts + reasons for non-acceptance (must be persuasive)
   - Reject: Provide sufficient scholarly rationale (literature, data, methodological argumentation)

4. Risk assessment
   "If you reject this suggestion, what might the reviewer's reaction be? Is it worth the risk?"

Key principles:

• Do not default to "accept all": Real reviewer comments are not always correct — some may be based on misunderstanding or school-of-thought bias
• 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?"
• User can say "just fix it for me" to skip: But when skipping strategic discussion, AI defaults to accepting all comments (conservative strategy)
• Maximum 8 rounds of dialogue, but at least 1 round per Major comment

Step 3: Revision and Response to Reviewers

Produce two documents:

1. Revised draft
   - Track all modification locations (additions/deletions/rewrites)
   - Revision content consistent with Response to Reviewers

2. Response to Reviewers letter
   Format (point-by-point response):
   +------------------------------------+
   | Reviewer [N], Comment [M]:         |
   |                                    |
   | [Original comment quote]           |
   |                                    |
   | Response:                          |
   | [Response explanation]             |
   |                                    |
   | Changes made:                      |
   | [Specific modification location    |
   |  and content]                      |
   | (or: We respectfully disagree      |
   |  because... [rationale])           |
   +------------------------------------+

Step 4: Self-Verification (Completeness Check)

Stage 3' behavior adjustments in external review mode:

1. Point-by-point comparison of External Review Summary with Response to Reviewers:
   - Does every comment have a response? (completeness)
   - Is each response consistent with actual changes? (consistency)
   - Were the places claimed as "modified" actually changed? (truthfulness)

2. New citation verification:
   - New references added during revision enter Stage 4.5 integrity verification

3. Things NOT done (different from internal review):
   - Do not reassess paper quality (that is the human reviewers' job)
   - Do not issue a new Editorial Decision
   - Do not raise new revision requests

Honest Capability Boundaries

1. 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
2. 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
3. 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
4. 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

Progress Dashboard

Users can say "status" or "pipeline status" at any time to view:

+=============================================+
|   Academic Pipeline v2.0 Status             |
+=============================================+
| Topic: Impact of AI on Higher Education     |
|        Quality Assurance                    |
+---------------------------------------------+

  Stage 1   RESEARCH          [v] Completed
  Stage 2   WRITE             [v] Completed
  Stage 2.5 INTEGRITY         [v] PASS (62/62 refs verified)
  Stage 3   REVIEW (1st)      [v] Major Revision (5 items)
  Stage 4   REVISE            [v] Completed (5/5 addressed)
  Stage 3'  RE-REVIEW (2nd)   [v] Accept
  Stage 4'  RE-REVISE         [-] Skipped (Accept)
  Stage 4.5 FINAL INTEGRITY   [..] In Progress
  Stage 5   FINALIZE          [ ] Pending
  Stage 6   PROCESS SUMMARY   [ ] Pending

+---------------------------------------------+
| Integrity Verification:                     |
|   Pre-review:  PASS (0 issues)              |
|   Final:       In progress...               |
+---------------------------------------------+
| Review History:                             |
|   Round 1: Major Revision (5 required)      |
|   Round 2: Accept                           |
+=============================================+

See templates/pipeline_status_template.md for the output template.

Revision Loop Management

• Stage 3 (first review) -> Stage 4 (revision) -> Stage 3' (verification review) -> Stage 4' (re-revision, if needed) -> Stage 4.5 (final verification)
• Maximum 1 round of RE-REVISE (Stage 4'): If Stage 3' gives Major, enter Stage 4' for revision then proceed directly to Stage 4.5 (no return to review)
• Pipeline overrides academic-paper's max 2 revision rule: In the pipeline, revisions are limited to Stage 4 + Stage 4' (one round each), replacing academic-paper's max 2 rounds rule
• Mark unresolved issues as Acknowledged Limitations
• Provide cumulative revision history (each round's decision, items addressed, unresolved items)

Reproducibility

v2.0 design ensures consistent quality assurance with each execution:

Standardized Workflow

Audit Trail

When the pipeline ends, state_tracker_agent produces a complete audit trail:

Pipeline Audit Trail
====================
Topic: [topic]
Started: [time]
Completed: [time]
Total Stages: [X/9]

Stage 1 RESEARCH: [mode] -> [output count]
Stage 2 WRITE: [mode] -> [word count]
Stage 2.5 INTEGRITY: [PASS/FAIL] -> [refs verified] / [issues found -> fixed]
Stage 3 REVIEW: [decision] -> [items count]
Stage 4 REVISE: [items addressed / total]
Stage 3' RE-REVIEW: [decision]
Stage 4' RE-REVISE: [executed / skipped]
Stage 4.5 FINAL INTEGRITY: [PASS/FAIL] -> [refs verified]
Stage 5 FINALIZE: Ask format style -> MD + DOCX + LaTeX (apa7/ieee/etc.) -> tectonic -> PDF
Stage 6 PROCESS SUMMARY: Ask language -> MD -> LaTeX -> PDF (zh/en)

Integrity Summary:
  Pre-review: [X] refs checked, [Y] issues found, [Y] fixed
  Final: [X] refs checked, [Y] issues found, [Y] fixed
  Overall: [CLEAN / ISSUES NOTED]

Stage 6: Process Summary Protocol (Added in v2.4)

Trigger: After Stage 5 (FINALIZE) completion Purpose: Document the complete human-AI collaboration history for the paper creation process, for user sharing, reporting, or reflection

Workflow

1. Ask user language preference:
   "Which language version of the process record would you like to generate first?"
   - Chinese (Traditional Chinese)
   - English
   - Both (default: generate the user's primary conversation language first)

2. Review session history and compile the following:
   - User's initial instructions (verbatim quote)
   - Key decision points and user interventions at each stage
   - Direction correction moments and reasons
   - Iteration count and review result summaries
   - Intellectual insights raised by the user (e.g., questions that spawned new chapters)
   - Quality requirement evolution (e.g., formatting, tone adjustments)
   - Pipeline statistics (stage count, review rounds, integrity verification count, etc.)

3. Generate Markdown version (paper_creation_process.md / paper_creation_process_en.md)

4. Convert to LaTeX and compile PDF:
   - pandoc MD -> LaTeX body
   - Package complete LaTeX document (with cover page, table of contents, headers/footers)
   - tectonic compile PDF
   - Chinese version requires xeCJK + Source Han Serif TC VF

Required Content in Process Record

Collaboration Quality Evaluation (Final Chapter, Mandatory)

The 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.

Scoring Dimensions (each 1-100, weighted average for overall score)

+--------------------------------------------------+
|  Collaboration Quality Score: [XX]/100            |
+--------------------------------------------------+
|                                                   |
|  Direction Setting          [----------  ] XX     |
|  Clarity, timing, scope definition                |
|                                                   |
|  Intellectual Contribution  [------------ ] XX    |
|  Insight depth, original questions, concept        |
|  challenges                                       |
|                                                   |
|  Quality Gatekeeping        [---------   ] XX     |
|  Visual inspection, formatting requirements,       |
|  quality standards                                |
|                                                   |
|  Iteration Discipline       [----------  ] XX     |
|  Timely direction correction, willingness to       |
|  re-run pipeline, refusing to settle              |
|                                                   |
|  Delegation Efficiency      [-------     ] XX     |
|  When to intervene/when to let go, instruction     |
|  precision, checkpoint efficiency                 |
|                                                   |
|  Meta-Learning              [------------ ] XX    |
|  Feeding experience back to skills, requesting     |
|  lesson recording, process improvement awareness  |
|                                                   |
+--------------------------------------------------+

Scoring Criteria

Required Subsections

1. Overall Score: Total score + one-sentence evaluation
2. What Worked Well: 2-4 specific behaviors, with verbatim user quotes
3. Missed Opportunities: 1-3 things the user could have done but didn't
4. Recommendations for Next Time: 3-5 specific, actionable improvement suggestions
5. Human vs AI Value-Add: Clearly identify which aspects of the final paper quality came from user intervention (not achievable by AI independently)

Evaluation Principles

• Honesty first: No inflation, no pleasantries. If the user only pressed "continue," reflect that truthfully
• Evidence-based: Every score is supported by specific behaviors or conversation records
• Constructive: Every criticism must include actionable improvement suggestions
• Acknowledge uncertainty: If certain dimensions cannot be evaluated (e.g., mid-entry skipped the research stage), mark as N/A
• Bidirectional reflection: Also candidly point out Claude's shortcomings during the process (e.g., areas requiring multiple corrections)

Output Specifications

• Filename: paper_creation_process.md (Chinese) / paper_creation_process_en.md (English)
• PDF: paper_creation_process_zh.pdf / paper_creation_process_en.pdf
• LaTeX template: article class, 12pt, A4, Times New Roman + Source Han Serif TC VF
• Includes table of contents: \tableofcontents
• Header: left = document title (italic), right = date
• Compilation: tectonic (same toolchain as Stage 5)

Quality Standards

Error Recovery

Agent File References

Reference Files

Templates

Examples

Output Language

Follows user language. Academic terminology retained in English.

Integration with Other Skills

academic-pipeline dispatches the following skills (does not do work itself):

Stage 1: deep-research
  - socratic mode: Guided research exploration
  - full mode: Complete research report
  - quick mode: Quick research summary

Stage 2: academic-paper
  - plan mode: Socratic chapter-by-chapter guidance
  - full mode: Complete paper writing

Stage 2.5: integrity_verification_agent (Mode 1: pre-review)
Stage 4.5: integrity_verification_agent (Mode 2: final-check)

Stage 3: academic-paper-reviewer
  - full mode: Complete 5-person review (EIC + R1/R2/R3 + Devil's Advocate)

Stage 3': academic-paper-reviewer
  - re-review mode: Verification review (focused on revision responses)

Stage 4/4': academic-paper (revision mode)
Stage 5: academic-paper (format-convert mode)
  - Step 1: Ask user which academic formatting style (APA 7.0 / Chicago / IEEE, etc.)
  - Step 2: Auto-produce MD + DOCX
  - Step 3: Produce LaTeX (using corresponding document class, e.g., apa7 class for APA 7.0)
  - Step 4: After user confirms content is correct, tectonic compiles PDF (final version)
  - Fonts: Times New Roman (English) + Source Han Serif TC VF (Chinese) + Courier New (monospace)
  - PDF must be compiled from LaTeX (HTML-to-PDF is prohibited)

Related Skills

Version Info

Changelog