experiment-execution

name	experiment-execution
description	Use when G2 (plan freeze) and G3 (execution readiness) gates have passed — handles all implementation, baseline reproduction, experimentation, and iteration for both Type M and Type D projects; this is Phase 4

Do NOT begin full-scale experiments without passing G3. Smoke tests are permitted before G3.

Experiment Execution (Phase 4)

Overview

This skill orchestrates all implementation and experimentation. It operates in TWO stages:

• Phase 4a (Exploratory): Quick, focused exploration to validate assumptions and discover insights BEFORE committing to full-scale execution. This stage may reveal that the research question or method design needs refinement.
• Phase 4b (Full Execution): Complete, rigorous execution of the (possibly refined) plan.

This two-stage model reflects how real research works: you don't know everything before you start, and early exploration often reshapes the direction.

On-Demand Literature Search (Active Throughout Phase 4)

Literature retrieval is NOT limited to Phase 1. During experiment execution, when you encounter unexpected results, need to understand a technique, or want to compare with a method not in the original literature review — search immediately. Use web search, arXiv, Google Scholar. Add new papers to docs/02_literature/paper-list.md with tag [Found during Phase 4].

If search fails or returns insufficient results:

1. Try alternative search terms (synonyms, broader/narrower scope, different field terminology)
2. If still insufficient → note what was searched and what's missing in docs/02_literature/paper-list.md
3. Ask the user: "I need papers on [topic] but couldn't find them via [methods tried]. Can you point me to relevant work?"
4. Do NOT block on literature — proceed with available information and mark the gap as [literature pending]

Global Constraints

Inject ALL of these into every subagent dispatch during Phase 4:

1. VALUE_ANCHOR: "This project's value proposition is: [read value_proposition from docs/01_intake/research-anchor.yaml]. All work serves this."
2. METRIC_LOCK: "Primary metrics are [read primary_metrics from docs/03_plan/evaluation-protocol.yaml]. IMMUTABLE without user permission."
3. VENUE_ALIGNMENT: "Target venue: [read target_venue.primary from research-anchor.yaml]. Is current experiment scale adequate?"
4. ANTI_CHERRY_PICK: "All seeds. All results. All failures recorded."
5. PREMATURE_EXPERIMENT_GUARD: "Baselines must be reproduced before running own method at full scale."

G3 Gate Checklist — Execution Readiness

Before entering Phase 4a, verify ALL items:

• Data arrived and validated (docs/04_data_resource/data-quality-report.md exists)
• [ML] Leakage audit passed (docs/04_data_resource/leakage-audit.md clean)
• Resources confirmed sufficient (compute matches research-anchor.yaml resources.compute)
• Tool chain ready (environment reproducible, dependencies pinned)
• Baseline implementation plan confirmed (sources listed in evaluation-protocol.yaml)

If ANY item fails, STOP and resolve before proceeding.

Execution Flow (Two-Stage)

digraph execution {
    rankdir=TB;
    g2 [label="G2 passed\nPlan frozen" shape=doublecircle];
    g3 [label="G3 check\nExecution readiness" shape=diamond];
    g3_fail [label="Resolve blockers" shape=box style=filled fillcolor="#f8d7da"];
    
    subgraph cluster_4a {
        label="Phase 4a — Exploratory";
        style=dashed;
        color=blue;
        exp_M [label="Type M: Run 1-2 baselines\n+ error analysis" shape=box style=filled fillcolor="#e0e0ff"];
        exp_D [label="Type D: Exploratory data analysis\n+ pattern discovery" shape=box style=filled fillcolor="#cce5ff"];
        exp_T [label="(Optional) Theoretical\nanalysis / proof sketch" shape=box style=filled fillcolor="#f0e0ff"];
    }
    
    decision [label="Decision Point:\nDoes plan still\nmake sense?" shape=diamond style=filled fillcolor="#fff3cd"];
    return_p2 [label="Return to Phase 2\n(refine question)" shape=box style=filled fillcolor="#ffcccc"];
    return_p3 [label="Return to Phase 3\n(refine design)" shape=box style=filled fillcolor="#ffcccc"];
    adjust [label="Adjust plan\n+ continue" shape=box style=filled fillcolor="#ffffcc"];
    
    subgraph cluster_4b {
        label="Phase 4b — Full Execution";
        style=dashed;
        color=green;
        A [label="Stage A\nBuild framework" shape=box];
        B [label="Stage B\nReproduce baselines" shape=box style=filled fillcolor="#d4edda"];
        C [label="Stage C\nImplement method\n(+ optional proofs)" shape=box];
        D [label="Stage D\nIterate & validate" shape=box];
        pivot [label="pivot-or-kill\n(after 3 failures)" shape=doubleoctagon style=filled fillcolor="#f8d7da"];
        E [label="Stage E\nFull-scale experiments" shape=box style=filled fillcolor="#d4edda"];
    }
    
    results [label="Phase 5:\nresults-integration" shape=doublecircle];

    g2 -> g3;
    g3 -> g3_fail [label="fail"];
    g3_fail -> g3 [label="retry"];
    g3 -> exp_M [label="pass\nType M"];
    g3 -> exp_D [label="pass\nType D"];
    exp_M -> exp_T [label="has theory?"];
    exp_D -> exp_T [label="has theory?"];
    exp_M -> decision;
    exp_D -> decision;
    exp_T -> decision;
    
    decision -> return_p2 [label="question\nwrong"];
    decision -> return_p3 [label="design\nneeds change"];
    decision -> adjust [label="minor\nadjustment"];
    decision -> A [label="plan\nconfirmed"];
    adjust -> A;
    
    A -> B;
    B -> C;
    C -> D;
    D -> D [label="iterate"];
    D -> pivot [label="3 failures"];
    pivot -> D [label="pivot approved"];
    D -> E [label="user confirms\nmethod works"];
    E -> results;
}

---

Phase 4a — Exploratory Stage

Phase 4a is NOT optional. Before committing to full-scale execution, you MUST do a focused exploration to validate that the plan from Phase 3 is sound. This catches fundamental problems BEFORE weeks of computation are wasted.

Phase 4a is deliberately QUICK — it should take a fraction of the effort of full execution. The goal is insight, not completeness.

Phase 4a for Type M — Baseline Probing + Initial Assessment

Goal: Understand the actual problem landscape before committing to full-scale execution.

NOTE: Baselines run here count toward Phase 4b Stage B — they will NOT be re-run. Phase 4a is a subset of the full execution, not a separate exercise.

1. Set up minimal infrastructure — data loading, evaluation function, seed control
2. Run 1-2 strongest baselines on 1 representative dataset, full execution (not toy run)
3. Analyze baseline results:
   • What's the actual performance level? Does it match literature?
   • Where do baselines fail? (error analysis, failure cases, edge cases)
   • Is there a systematic failure pattern? Is it what we predicted in Phase 3?
4. Quick prototype of our method (optional but recommended) — run a minimal version to see if the core idea shows promise
5. Quick literature check if findings are surprising

Output: docs/05_execution/phase4a-exploration-report.md

Phase 4a Exploration Report (Type M):
═════════════════════════════════════

Baselines run: [list with results]

Error analysis findings:
  [describe failure patterns, edge cases, systematic issues]

Predicted failure pattern (from Phase 3): [X]
Actual failure pattern: [Y]
Alignment: [matched / partially matched / different]

Quick prototype results (if run): [summary]

Issues or surprises: [list anything unexpected]

Phase 4a for Type D — Preliminary Analysis Run

Goal: Run the core analysis pipeline once on real data to see what the actual results look like, before committing to the full analysis plan.

NOTE: "Exploratory analysis" means different things in different fields. It is NOT limited to clustering and dimensionality reduction. The point is to execute the main planned analysis at a basic level and examine the output.

1. Preprocessing — quality control, normalization, whatever the domain requires
2. Run the core analysis pipeline once — whatever the Phase 3 plan calls for as the main analysis (could be differential expression, could be model fitting, could be statistical tests, could be simulation — depends on the field)
3. Examine the results:
   • Do the results make sense given domain knowledge?
   • Are the expected signals present?
   • Is there something unexpected or interesting?
   • Is the data quality sufficient for the planned analysis?
4. Assess the analysis plan:
   • Does the Phase 3 plan still make sense given what we see?
   • Are some planned analyses unnecessary? Are there gaps?

Output: docs/05_execution/phase4a-exploration-report.md

Phase 4a Exploration Report (Type D):
═════════════════════════════════════

Data quality: [summary]

Preliminary analysis results:
  [describe key outputs from the initial analysis run]

Results vs expectations:
  Expected: [what Phase 3 predicted]
  Observed: [what actually came out]
  Surprises: [anything unexpected]

Assessment of analysis plan:
  Still appropriate? [yes / needs adjustment]
  Missing analyses? [list]
  Unnecessary analyses? [list]

Phase 4a for Type C — Quick Benchmark Probe

Goal: Run a quick comparison on one benchmark to see if the tool's advantages are real.

1. Run our tool + 1-2 main competitors on 1 representative benchmark
2. Compare: correctness, speed, memory
3. Assess: Does our claimed advantage hold? How big is the gap?
4. Test installation: Does it install cleanly on a fresh environment?

Output: Append to docs/05_execution/phase4a-exploration-report.md

Phase 4a Exploration Report (Type C):
═════════════════════════════════════

Quick benchmark: [name]
Our tool: [correctness, runtime, memory]
Competitor A: [correctness, runtime, memory]
Competitor B: [correctness, runtime, memory]

Claimed advantage: [X]
Actual advantage: [Y]
Gap: [large / moderate / marginal / none]

Installation test: [clean / issues]

Phase 4a — Optional: Theoretical Analysis / Proof Sketch

If the project includes theoretical claims (convergence, complexity, correctness, bounds):

1. Sketch the proof strategy — identify the key lemmas needed
2. Attempt the easiest lemma first — to check if the overall approach is viable
3. Identify blockers — are there steps that require assumptions we can't justify?
4. Assess feasibility — can we prove what we claimed, or do we need to weaken the claim?

This can run in parallel with the computational exploration above.

Output: Append to docs/05_execution/phase4a-exploration-report.md

Theoretical Analysis (preliminary):
  Claim: [what we want to prove]
  Proof strategy: [approach]
  Key lemma attempted: [which one, result]
  Feasibility: [viable / needs weaker claim / blocked by X]

⛔ Phase 4a Decision Point (MANDATORY STOP)

After Phase 4a, you MUST present the exploration report to the user and STOP. The user decides what to do next. Do NOT automatically proceed to Phase 4b.

Based on the exploration results, assess the situation and present appropriate options. The most common outcome is local adjustment, not a full phase reset.

Phase 4a exploration complete. Report above.

MY ASSESSMENT: [summarize what was found and what it means for the plan]

YOUR OPTIONS:

1. ✅ PROCEED AS-IS — Plan is confirmed. Continue to Phase 4b.

2. 🔧 LOCAL ADJUSTMENT (most common) — Adjust specific parts of the 
   plan based on what we learned:
   [list specific proposed changes, e.g.:]
   - Add experiment X to address [finding]
   - Modify analysis step Y because [reason]
   - Replace baseline Z with a stronger one
   - Adjust method component W based on error analysis
   These changes are applied directly — no need to redo Phase 3.

3. 🔄 FOCUSED RE-DISCUSSION — Some findings need deeper thought. I'll 
   dispatch the Phase 3 agents (Innovation Advisor / Domain Scientist / etc.)
   for a FOCUSED discussion on the specific issue:
   "[describe the specific issue that needs multi-agent input]"
   This is NOT a full Phase 3 redo — just a targeted consultation.

4. ↩️ RETURN TO PHASE 3 — The method/analysis design has a FUNDAMENTAL 
   problem that can't be fixed with local adjustments.
   [Only suggest this if there's a deep structural issue.]

5. ↩️↩️ RETURN TO PHASE 2 — The research question itself is wrong.
   [Only suggest this in extreme cases — e.g., the data contradicts 
   the premise, or the problem doesn't exist.]

Which option? (You can also propose your own adjustments.)

Guidelines for which option to suggest:

Situation	Recommended Option
Results match predictions, plan looks good	1 (Proceed)
Baseline stronger/weaker than expected, need to add/change experiments	2 (Local Adjustment)
Error analysis reveals a different bottleneck than predicted	2 or 3 (Adjust or Re-discuss)
Data shows unexpected pattern worth exploring	2 (Add experiments to explore it)
Method's core idea doesn't address the actual problem	3 or 4 (Re-discuss or Return)
Data doesn't support the research question at all	5 (Return to Phase 2)

For Option 2 (Local Adjustment):

• Apply the changes directly to Phase 3 outputs (evaluation protocol, analysis plan, etc.)
• Document what changed and why in the exploration report
• Proceed to Phase 4b with the updated plan

For Option 3 (Focused Re-discussion):

• Dispatch the relevant Phase 3 agents with the Phase 4a findings
• Give them a FOCUSED prompt: "Based on these exploration results, should we [specific question]?"
• Agents see the exploration report + the original plan
• Apply their recommendations and proceed to Phase 4b
• This is a LIGHTWEIGHT version of the Phase 3 deliberation — max 2-3 rounds, focused on one issue

For Options 4 and 5 (Full return):

• Save exploration findings in docs/05_execution/phase4a-exploration-report.md
• Return to the appropriate phase via its "Re-entry from Phase 4a" path
• Carry forward all exploration insights

The user can also iterate within Phase 4a: If the user says "try X first" or "run one more experiment before deciding," that's fine — Phase 4a can have multiple rounds of quick exploration before the decision point. The MANDATORY STOP is about getting user input before committing to full-scale Phase 4b, not about limiting exploration.

---

Phase 4b — Full Execution

---

Type M Execution Flow (Phase 4b)

Stage A — Build Experiment Framework

Set up the infrastructure before any experiment runs:

1. Data loading pipeline (scripted, deterministic)
2. Evaluation function matching evaluation-protocol.yaml exactly
3. Logging framework (metrics, configs, seeds, timestamps)
4. Reproducibility scaffolding: seed control, environment snapshot, config saving

Use implementer-prompt.md to dispatch coding tasks.

Stage B — Complete Baseline Reproduction

IRON LAW: BASELINES FIRST. NO EXCEPTIONS.

NOTE: If Phase 4a already ran some baselines, carry those results forward — do NOT re-run them. Stage B completes the REMAINING baselines and consolidates all results.

B.1 — Check for pre-computed and Phase 4a results

Before implementing anything, check two sources:

1. evaluation-protocol.yaml — each baseline's pre_computed.status
2. docs/05_execution/phase4a-exploration-report.md — baselines already run in Phase 4a

Status	Action
accepted	Skip re-running. Import pre-computed results directly into baseline-results.md. Verify the numbers match what was recorded.
reference_only	Must re-run under our evaluation protocol. Use pre-computed results as a sanity check — our reproduced numbers should be in the same ballpark.
incompatible	Must re-run from scratch. Pre-computed results are discarded for official comparison.
(empty)	Must run — no pre-computed results available.

Report to user:

Baseline execution plan:
  Skip (pre-computed accepted): N baselines
  Re-run (reference only / incompatible / new): M baselines
  Total baselines: N + M

B.2 — Run baselines that need running

For each baseline that needs running:

1. Implement or integrate from official source
2. Run with ALL seeds from evaluation-protocol.yaml
3. Compare with paper-reported values (or reference_only pre-computed values)

Discrepancy vs reference	Action
≤ 1% relative	Acceptable. Proceed.
1–5% relative	Investigate. Document reason. Proceed if justified.
> 5% relative	STOP. Debug until resolved or user approves deviation.

B.3 — Consolidate all baseline results

Merge accepted pre-computed results and newly reproduced results into a single table in docs/05_execution/baseline-results.md:

Baseline	Source	Dataset	Primary Metric	Seeds	Notes
Method A	accepted pre-computed (user's own run)	X	0.87 ± 0.02	5	Imported from Phase 1
Method B	reproduced (official repo)	X	0.83 ± 0.01	5	Matched paper within 0.5%

All downstream comparisons (Stage D, Stage E) use this consolidated table.

Stage C — Implement Core Method (+ Optional Theoretical Analysis)

Build incrementally — each component must be independently testable:

1. Implement smallest viable version first
2. Unit test each module
3. Follow reproducibility-driven-research: fix seeds, log environment, script everything
4. Commit after each working component

If the project includes theoretical claims (convergence, complexity, correctness):

Complete the formal proofs alongside or before full experimental validation:

1. State theorems formally — precise conditions, claims, and notation
2. Prove each theorem — show complete derivations, not hand-waving
3. Verify assumptions — ensure the conditions required by the proof hold in practice
4. Connect theory to experiments — the experiments should validate what the theory predicts

Write proofs to docs/05_execution/theoretical-analysis.md (or .tex for LaTeX).

Common proof types in computational research:

Type	Example	What to Show
Convergence	"Our optimizer converges to a stationary point"	Rate and conditions
Complexity	"Algorithm runs in O(n log n) time"	Tight bound with proof
Correctness	"Output satisfies property P"	Formal proof, possibly by induction
Approximation	"Solution is within factor α of optimal"	Bound with construction
Generalization	"Test error bounded by X with probability Y"	PAC/VC/Rademacher bound

If a proof attempt FAILS (can't prove the desired result):

• Document what went wrong — is the claim too strong?
• Consider weakening the theorem (e.g., weaker conditions, looser bound)
• Discuss with user: proceed with weaker theoretical guarantee, or redesign the method?

Stage D — Initial Experiments + Iteration

RUN TO COMPLETION. Every method must execute its full intended procedure before results can be reported. Partial execution results are NOT valid experimental results — they are at best smoke tests. What "full execution" means depends on the method type:

• Iterative optimization (deep learning, gradient descent): Train until convergence — validation metric stops improving for a patience window (defined in evaluation-protocol.yaml training_convergence section). 1-epoch or few-epoch results are NOT experiments.
• Non-iterative ML (random forest, SVM, KNN, etc.): Fit with the full training set and intended hyperparameters. A quick test on a tiny subset is NOT a final run.
• Simulation / optimization: Run until the convergence criterion defined in the experiment plan is satisfied (e.g., objective stabilizes, residual below threshold).
• Statistical analysis / pipeline: Execute the complete analysis pipeline as designed. Partial pipeline outputs are intermediate, not results.

Reporting incomplete execution as final results is FORBIDDEN.

Small-scale validation first. Do NOT run full experiments immediately.

1. Run on 1 dataset at reduced scale — but the method itself must still run to completion (e.g., train to convergence for DL, full fit for ML, full pipeline for analysis)
2. Compare with baselines from Stage B on the SAME dataset with the SAME execution regime
3. If improvement → expand to more datasets
4. If no improvement → enter the Mandatory Iteration Loop below

Mandatory Iteration Loop (Type M)

MINIMUM 3 ITERATION ROUNDS before declaring a method "doesn't work." Each round must follow the diagnose-hypothesize-fix-measure cycle below. Random changes without analysis are NOT iterations — they are noise.

For each iteration round:

Step	Action	Output
1. Diagnose	Analyze WHERE the method fails — error analysis, confusion matrices, loss curves, per-class/per-sample breakdown	Specific failure pattern
2. Hypothesize	State WHY based on diagnosis. One hypothesis per round.	Written hypothesis
3. Fix	Make ONE targeted change addressing the hypothesis	Code change + rationale
4. Measure	Run method to completion, evaluate on ALL seeds, compare to previous round AND baselines	Updated results table
5. Report	Present to user: "Round N: changed X because Y. Result: Z vs previous W vs baseline V."	User acknowledgment

After each round, present this comparison to the user:

=== Iteration Round N Report ===
Hypothesis: [what was wrong]
Change: [what was modified]
Result:
  This round:  XX.X ± Y.Y (primary metric)
  Last round:  XX.X ± Y.Y
  Best baseline: ZZ.Z ± W.W
  Gap to best baseline: ±N.N%
Status: [improving / stagnating / regressing]
Recommendation: [continue iterating / ready for full-scale / consider pivot]

User must acknowledge each report before the next round begins.

Minimum Performance Bar

Before moving from Stage D to Stage E, the method MUST satisfy:

• Hard minimum: Performance within 80% of the best baseline on primary metric (e.g., if baseline is 0.95 accuracy, method must reach at least 0.76)
• Expected: Method matches or exceeds at least one baseline on primary metric

If the method is drastically worse than all baselines (e.g., 0.42 vs 0.95):

1. This signals a fundamental problem, not a tuning issue
2. Do NOT proceed to Stage E
3. Do NOT proceed to results integration
4. Diagnose: did the method run to completion? Is there a bug? Are hyperparameters reasonable? Is the approach fundamentally flawed?
5. If 3 full iteration rounds fail to close the gap → trigger pivot-or-kill

After 3 consecutive failures → trigger pivot-or-kill. No fourth attempt without user decision.

Stage E — Full-Scale Experiments

Enter ONLY after user explicitly confirms the method works at small scale AND the minimum performance bar is met.

Run the complete experiment matrix:

1. All datasets × all methods × all seeds from evaluation-protocol.yaml
2. Ablation study per docs/03_plan/ablation-design.md
3. Parameter sensitivity for key hyperparameters
4. Efficiency analysis (runtime, memory, parameter count)
5. Domain-specific pitfall checks:

Domain	Pitfall Checks
ML	Overfitting (train vs test gap), data leakage, label noise sensitivity
Bioinformatics	Batch effects, sequence length bias, homology leakage
Physics	Conservation law violations, symmetry breaking, unit consistency
Chemistry	Energy conservation, chirality handling, force field artifacts

Dispatch methodology-reviewer-prompt.md after Stage E completes. Dispatch reproducibility-reviewer-prompt.md before declaring results final.

---

Type D Execution Flow (Phase 4b)

Stage A — Data Preprocessing + Quality Control

All preprocessing must be scripted and reproducible:

1. Raw data → processed data pipeline (raw data NEVER modified)
2. Quality control report with statistics
3. Missing data, outlier, and distribution analysis
4. Document every preprocessing decision

Stage B — Execute Analysis Framework

Follow the analysis framework from Phase 3, in order:

1. Descriptive statistics — distributions, summaries, initial visualizations
2. Main analysis — hypothesis testing or primary modeling
3. Mechanism exploration — why does the pattern exist?
4. Robustness checks — sensitivity, subgroups, alternative specifications
5. Exploratory analysis — unexpected patterns

For each step:

• Generate figures — invoke amplify:figure-quality-standards for every figure. Apply the style template and method-color mapping from the first figure onward.
• Self-check: "Are these results reasonable given domain knowledge?"
• Record in docs/05_execution/experiment-log.md

Stage C — Alternative Hypothesis Exclusion

REQUIRED SUB-SKILL: Invoke alternative-hypothesis-check if available.

For each main finding, systematically rule out alternative explanations identified in Phase 3 (docs/03_plan/analysis-storyboard.md). Document evidence for and against each alternative.

Stage D — Sufficiency Self-Check

Verify against Phase 3 sufficiency criteria:

• Data explored from multiple angles (not just one clustering/one visualization)?
• Main hypothesis tested from multiple perspectives?
• Exploratory analysis completed for unexpected patterns?
• All analyses connect back to story line?
• Statistical support sufficient for each claim?
• Confounders identified and addressed?
• At least 4–6 content points with figures/tables?
• At least 4 distinct, publication-quality figures generated?

If any item fails, return to the relevant stage before declaring Phase 4 complete.

Domain Sanity Check (All Types)

Before declaring ANY results valid, perform a **domain sanity check**: do the results make sense given basic domain knowledge? AI agents can produce technically correct but biologically/scientifically nonsensical results without noticing.

Examples of sanity failures (catch these BEFORE presenting results):

• PBMC analysis shows B cells at 59% and no T cells → T cells should be ~60-70% of PBMCs; this means annotations are wrong
• Gene expression analysis shows all housekeeping genes as top markers → likely a normalization problem
• Clustering produces 50 clusters from 2,700 cells → over-clustering
• All p-values are exactly 0.0 → likely a numerical issue, not real significance
• A "novel subpopulation" that contains 3 cells → too few to be reliable
• Trajectory analysis shows all cells on one branch → pseudotime may have failed

How to check:

1. Compare key quantitative results against known ranges from the literature
2. If a result seems surprising, verify it's genuinely novel rather than an artifact
3. If cell type proportions, effect sizes, or distributions differ dramatically from literature expectations, investigate before reporting
4. Ask: "Would a domain expert look at this result and immediately say something is wrong?"

If a sanity check fails, debug the analysis pipeline before continuing. Do NOT report suspicious results as "findings."

Type D Iteration Requirement

Type D projects also require iteration. "Ran analysis once and got some results" is NOT sufficient.

After initial analysis, ask yourself:

1. Are the findings interesting enough for the target venue? If not → deepen the analysis.
2. Are there sub-populations, edge cases, or unexpected patterns worth exploring? If yes → run additional analyses.
3. Does the story line from Phase 3 have gaps? If yes → fill them with additional analyses.
4. Are the figures compelling or just "we did the analysis"? If the latter → generate better visualizations.
5. Do the results pass the domain sanity check above? If not → debug before iterating.

Minimum: Present the user with at least 2 rounds of results before declaring Phase 4 complete. Round 1 = initial analysis. Round 2 = refined analysis addressing gaps, deepening insights, or exploring unexpected findings.

---

Type C Execution Flow (Phase 4b)

Stage A — Implementation + Testing

1. Core tool implementation (if not already done)
2. Comprehensive unit tests and integration tests
3. Edge case handling
4. Documentation: API reference, installation guide, quickstart tutorial
5. Package for distribution (pip, conda, CRAN, etc.)

Stage B — Benchmark Against Competitors

Run ALL competing tools and our tool on the SAME benchmark suite:

1. Correctness: Compare outputs against reference/gold standard
2. Performance: Runtime and memory on standardized inputs
3. Scalability: Scaling curves (vary input size systematically)

Record everything in docs/05_execution/benchmark-results.md.

Stage C — Real-World Case Studies

Run 2-3 real-world use cases from the target domain:

• Use realistic data, not toy examples
• Document the workflow end-to-end
• Show how the tool solves a real problem the user would have
• Generate figures showing results

Stage D — Usability and Documentation Review

• Test installation on a clean environment
• Run the quickstart tutorial as a new user would
• Check API consistency and error messages
• Verify documentation completeness

---

Subagent Integration

Use the Task tool to dispatch subagents for discrete implementation tasks (coding, running experiments, generating figures). Each subagent gets a focused task with clear constraints.

How to dispatch

Use prompt templates in this skill's directory. Fill in all bracketed fields, then call the Task tool:

Call Task tool with:
  description: "[short task name]"
  prompt: |
    [Contents of the filled-in prompt template]
  subagent_type: "generalPurpose"

Prompt templates (all in this skill's directory)

Template	When to dispatch	What it checks
implementer-prompt.md	Each coding/experiment task	Implements, runs to completion, records results
experiment-spec-reviewer-prompt.md	After each implementation	Did it do what was planned? Are results from complete runs? All seeds ran?
methodology-reviewer-prompt.md	After Stage B (baselines) and Stage E (full experiments)	Statistical rigor, baseline fairness, data integrity
reproducibility-reviewer-prompt.md	Once, before declaring Phase 4 complete	Can a third party reproduce everything?

Review loop per task

1. Dispatch implementer (Task tool) → implementer reports back with results
2. Dispatch spec reviewer (Task tool) → check completeness and compliance
3. If spec reviewer finds issues → dispatch implementer again with fix instructions → dispatch spec reviewer again
4. After major milestones (baselines done, full experiments done) → dispatch methodology reviewer (Task tool, using methodology-reviewer agent)
5. After ALL tasks done → dispatch reproducibility reviewer (Task tool, using reproducibility-reviewer agent)

Parallel dispatch within a stage

For truly independent tasks within the same stage (e.g., 3 independent baselines), call multiple Task tools in the same message:

Call Task tool with:
  description: "Run baseline: GraphSAGE"
  prompt: |
    [Filled implementer-prompt.md for GraphSAGE]
  subagent_type: "generalPurpose"

Call Task tool with:
  description: "Run baseline: GAT"
  prompt: |
    [Filled implementer-prompt.md for GAT]
  subagent_type: "generalPurpose"

Call Task tool with:
  description: "Run baseline: GIN"
  prompt: |
    [Filled implementer-prompt.md for GIN]
  subagent_type: "generalPurpose"

Important: Research stages (A→B→C→D→E) are strongly interdependent. Do NOT dispatch multiple stages in parallel. Run them sequentially, with review gates between major milestones. Only use parallel dispatch for independent work within a stage.

Red Flags — STOP

• Running full-scale experiments before baselines are reproduced
• Skipping small-scale validation and going straight to full experiments
• Attempting a 4th iteration after 3 consecutive failures without user consultation
• Changing primary metrics mid-experiment without user authorization
• Reporting partial seed results as complete
• Running own method with better hyperparameter budget than baselines

Rationalization Prevention

Excuse	Reality
"Baselines are slow, let me skip to my method"	Baselines first. No exceptions. Skipping guarantees unfair comparison.
"One seed is enough for now"	One seed is noise, not signal. Run all seeds or don't report.
"Let me try a completely different approach"	Analyze what failed first. Random pivots waste everything learned.
"Full-scale is the only real test"	Small-scale catches fatal flaws cheaply. Full-scale confirms, not discovers.
"The baseline paper used different settings"	Match their settings first. Then run your settings. Report both.
"Negative results slow us down"	Negative results are data. Record them. They prevent repeating mistakes.

Checklist

1. Verify G3 gate checklist passes
2. Branch to correct execution flow (Type M / Type D)
3. Complete all stages in order — no skipping
4. Dispatch methodology review after experiments
5. Dispatch reproducibility review before declaring results final
6. Record everything in docs/05_execution/

---

Re-Entry from Phase 5 (Experiment Supplement Path)

When Phase 5 (results integration) identifies missing experiments — typically flagged by the Devil's Advocate or Audience Specialist as REQUIRED supplements — the workflow returns here.

What you receive: A supplement request from Phase 5 specifying:

• Exactly which experiments/analyses are needed
• Why they are needed (which claim lacks evidence)
• Priority (REQUIRED vs RECOMMENDED)

What you do:

1. Scope check — Only run the requested supplements. Do NOT re-run existing experiments or re-open completed stages.
2. Follow the same discipline — Same seeds, same evaluation protocol, same fairness constraints. The locked protocol still applies.
3. Record in experiment log — Mark entries as [Supplement from Phase 5] in docs/05_execution/experiment-log.md.
4. Domain sanity check — Verify new results make scientific sense, same as for original experiments.
5. Return to Phase 5 — Present supplement results to the user, then resume Phase 5 where it left off. The multi-agent discussion panel will re-evaluate with the updated results.

Do NOT: Re-run the entire Phase 4 pipeline. Do NOT change the evaluation protocol. Do NOT add experiments beyond what was requested without user approval.

---

Re-Entry from Phase 6 (Paper Writing Supplement Path)

When Phase 6 (paper writing) review panel identifies that new experiments are needed — typically flagged as "EXPERIMENT NEEDED" by domain critics or adversarial reviewers — the workflow returns here.

What you receive: A supplement request from Phase 6 specifying:

• Which experiments/analyses are needed
• Which paper section or claim triggered the request
• Priority level

What you do:

1. Same scoped approach as Phase 5 supplements — Only run what's requested.
2. Record as [Supplement from Phase 6] in the experiment log.
3. Domain sanity check on new results.
4. Return to Phase 6 — Present supplement results to the user, then resume paper writing. The affected sections will be updated with the new results, and the polishing cycle for those sections will re-run.

Important: If the supplement changes results significantly (e.g., new findings that alter the narrative), the argument blueprint (argument-blueprint.md) should be updated before resuming Phase 6.

---

## ⛔ MANDATORY STOP — After Phase 4 Completion

When all experiments/analyses are done, END YOUR RESPONSE with a results summary:

Phase 4 complete. Results summary:
  - [list key results, metrics, figures generated]
  - [list any issues or limitations found]

Shall I proceed to Phase 5 (Results Integration)?
This will include a multi-agent discussion panel to stress-test
the results before paper writing.

Do NOT invoke results-integration or begin organizing results in the same response. Do NOT skip to paper writing.

STOP. WAIT. The user must reply before you proceed.