Research Pipeline

💡 Override via argument, e.g., /research-pipeline "topic" — AUTO_PROCEED: false, human checkpoint: true.

Full Autonomy Principle

This pipeline is designed to run fully autonomously without human intervention. At every decision point:

1. Never block waiting for user input — make the best decision based on available data, document the reasoning, and continue.
2. Auto-select at forks — when multiple options exist (ideas, variants, fix strategies), apply quantitative criteria to select the best one. Log the decision and alternatives.
3. Auto-recover from failures — when experiments fail, web searches hang, or reviews are harsh, diagnose the issue, apply the best fix, and continue. Exhaust at least 2 approaches before flagging as unresolvable.
4. Auto-infer missing context — when required files are absent (RESEARCH_BRIEF.md, EXPERIMENT_PLAN.md), infer the needed information from whatever IS available (CLAUDE.md, existing code, prior outputs).
5. Document all autonomous decisions — every auto-decision is logged with reasoning so the user can review after the fact. Format: [AUTO-DECISION] Chose X over Y because Z.

All downstream skills inherit this principle. No sub-skill should stop the pipeline to ask the user a question unless there is genuinely zero context to make any decision.

Overview

This skill chains the entire research lifecycle into a single pipeline:

/idea-discovery → implement → /run-experiment → /auto-review-loop → submission-ready
├── Workflow 1 ──┤            ├────────── Workflow 2 ──────────────┤

It orchestrates two major workflows plus the implementation bridge between them.

Pipeline

Stage 1: Idea Discovery (Workflow 1)

If RESEARCH_BRIEF.md exists in the project root, it will be automatically loaded as detailed context (replaces one-line prompt). See templates/RESEARCH_BRIEF_TEMPLATE.md.

Invoke the idea discovery pipeline:

/idea-discovery "$ARGUMENTS"

This internally runs: /research-lit → /idea-creator → /novelty-check → /research-review

Output: IDEA_REPORT.md with ranked, validated, pilot-tested ideas.

🚦 Gate 1 — Human Checkpoint:

After IDEA_REPORT.md is generated, pause and present the top ideas to the user:

📋 Idea Discovery complete. Top ideas:

1. [Idea 1 title] — Pilot: POSITIVE (+X%), Novelty: CONFIRMED
2. [Idea 2 title] — Pilot: WEAK POSITIVE (+Y%), Novelty: CONFIRMED
3. [Idea 3 title] — Pilot: NEGATIVE, eliminated

Recommended: Idea 1. Shall I proceed with implementation?

If AUTO_PROCEED=false: Wait for user confirmation before continuing. The user may:

• Approve an idea → proceed to Stage 2.
• Pick a different idea → proceed with their choice.
• Request changes (e.g., "combine Idea 1 and 3", "focus more on X") → update the idea prompt with user feedback, re-run /idea-discovery with refined constraints, and present again.
• Reject all ideas → collect feedback on what's missing, re-run Stage 1 with adjusted research direction. Repeat until the user commits to an idea.
• Stop here → save current state to IDEA_REPORT.md for future reference.

If AUTO_PROCEED=true: Present the top ideas, wait 10 seconds for user input. If no response, auto-select the #1 ranked idea (highest pilot signal + novelty confirmed) and proceed to Stage 2. Log: "AUTO_PROCEED: selected Idea 1 — [title]".

⚠️ This gate waits for user confirmation when AUTO_PROCEED=false. When true, it auto-selects the top idea after presenting results. The rest of the pipeline (Stages 2-4) is expensive (GPU time + multiple review rounds), so set AUTO_PROCEED=false if you want to manually choose which idea to pursue.

Stage 2: Implementation

Once the user confirms which idea to pursue:

1. Read the idea details from IDEA_REPORT.md (hypothesis, experimental design, pilot code)

2. Implement the full experiment:

   • Extend pilot code to full scale (multi-seed, full dataset, proper baselines)
   • Add proper evaluation metrics and logging (wandb if configured)
   • Write clean, reproducible experiment scripts
   • Follow existing codebase conventions

3. Code review: Before deploying, do a self-review:

   • Are all hyperparameters configurable via argparse?
   • Is the random seed fixed and controllable?
   • Are results saved to JSON/CSV for later analysis?
   • Is there proper logging for debugging?

Stage 3: Deploy Experiments (Workflow 2 — Part 1)

Deploy the full-scale experiments:

/run-experiment [experiment command]

What this does:

• Check GPU availability on configured servers
• Sync code to remote server
• Launch experiments in screen sessions with proper CUDA_VISIBLE_DEVICES
• Verify experiments started successfully

Monitor progress:

/monitor-experiment [server]

Wait for experiments to complete. Collect results.

Stage 4: Method Evolution (Workflow 2 — Part 2)

Once initial results are in, start the improvement loop.

If DEEP_INNOVATION = true (for projects requiring genuine methodological innovation):

/deep-innovation-loop "$ARGUMENTS — [chosen idea title]" — baseline: [PRIMARY_BASELINE], venue: [VENUE]

What this does (40+ rounds):

1. GPT-5.4 xhigh diagnoses root causes (not just symptoms)
2. Claude Code researches literature for techniques addressing root causes
3. GPT-5.4 proposes innovative method variants with "1+1>2" fusion design
4. Claude Code implements the best variant, runs experiments
5. Both reflect on results, update technique library and evolution log
6. Repeat until score ≥ 8/10 or convergence plateau

Output: innovation-logs/FINAL_METHOD.md, innovation-logs/EVOLUTION_LOG.md, innovation-logs/TECHNIQUE_LIBRARY.md

After deep-innovation-loop completes, optionally run /auto-review-loop for 2-3 rounds of final paper-level polish.

If DEEP_INNOVATION = false (default — quick iterative polishing):

/auto-review-loop "$ARGUMENTS — [chosen idea title]"

What this does (up to 4 rounds):

1. GPT-5.4 xhigh reviews the work (score, weaknesses, minimum fixes)
2. Claude Code implements fixes (code changes, new experiments, reframing)
3. Deploy fixes, collect new results
4. Re-review → repeat until score ≥ 6/10 or 4 rounds reached

Output: AUTO_REVIEW.md with full review history and final assessment.

Stage 5: Final Summary

After the auto-review loop completes, write a final status report:

# Research Pipeline Report

**Direction**: $ARGUMENTS
**Chosen Idea**: [title]
**Date**: [start] → [end]
**Pipeline**: idea-discovery → implement → run-experiment → auto-review-loop

## Journey Summary
- Ideas generated: X → filtered to Y → piloted Z → chose 1
- Implementation: [brief description of what was built]
- Experiments: [number of GPU experiments, total compute time]
- Review rounds: N/4, final score: X/10
- [If DEEP_INNOVATION=true] Innovation rounds: N, method evolution: v0 → vN, techniques explored: M, final vs baseline improvement: [metrics]

## Final Status
- [ ] Ready for submission / [ ] Needs manual follow-up

## Remaining TODOs (if any)
- [items flagged by reviewer that weren't addressed]

## Files Changed
- [list of key files created/modified]

Key Rules

• Large file handling: If the Write tool fails due to file size, immediately retry using Bash (cat << 'EOF' > file) to write in chunks. Do NOT ask the user for permission — just do it silently.

• Human checkpoint after Stage 1 is controlled by AUTO_PROCEED. When false, do not proceed without user confirmation. When true, auto-select the top idea after presenting results.

• Stages 2-4 can run autonomously once the user confirms the idea. This is the "sleep and wake up to results" part.

• If Stage 4 ends at round 4 without positive assessment, stop and report remaining issues. Do not loop forever.

• Budget awareness: Track total GPU-hours across the pipeline. Flag if approaching user-defined limits.

• Documentation: Every stage updates its own output file. The full history should be self-contained.

• Fail gracefully: If any stage fails (no good ideas, experiments crash, review loop stuck), report clearly and suggest alternatives rather than forcing forward.

Typical Timeline

Sweet spot: Run Stage 1-2 in the evening, launch Stage 3-4 before bed, wake up to a reviewed paper.