Experiment Audit: Fresh-Agent Integrity Verification

Before each reviewer slot, follow Selectable Reviewers: resolve invocation → .aris/reviewer.json → Codex default (fresh native reviewer) with reviewer_route.py through the canonical helper resolver, preserving this slot's tier, fresh contexts, source/image inputs and deterministic gates. Codex examples/pins below apply only to backend=codex; other routes use their own first-call/reply contract. Save the route and actual returned identity; unavailable selections are REVIEW_UNAVAILABLE. Classify acceptance from actual families/completion/scope. Forward invocation reviewer/model/effort overrides to reviewer-bearing children. Generation workers are unchanged.

Codex assurance: base semantic audit results record review_independence: same-family and acceptance_status: provisional. Deterministic evidence checks may be accepted; unavailable reviewer calls emit BLOCKED/ERROR rather than a provisional PASS.

Audit experiment integrity for: $ARGUMENTS

Why This Exists

LLM agents can produce fraudulent experimental results through:

1. Fake ground truth — creating synthetic "reference" from model outputs, then reporting high agreement as performance
2. Score normalization — dividing metrics by the model's own max to get 0.99+
3. Phantom results — claiming numbers from files that don't exist or functions never called
4. Insufficient scope — reporting 2-scene pilots as "comprehensive evaluation"

These are NOT intentional deception — they are failure modes of optimizing agents that lack integrity constraints. This skill adds that constraint.

Core Principle

The executor (Codex) collects file paths. A fresh Codex reviewer reads code and judges integrity. The executor does NOT participate in integrity judgment; this base route is same-family/provisional.

This follows shared-references/reviewer-independence.md and shared-references/experiment-integrity.md.

Constants

• REVIEWER_BACKEND = codex — Default: Codex reviewer agent (spawn_agent, ultra — deep-audit tier). Override with — reviewer: oracle-pro for GPT-5.5 Pro via Oracle MCP. See shared-references/reviewer-routing.md.

Workflow

Step 1: Collect Artifacts (Executor — Codex)

Locate and list these files WITHOUT reading or summarizing their content:

Scan project directory for:
1. Evaluation scripts:    *eval*.py, *metric*.py, *test*.py, *benchmark*.py
2. Result files:          *.json, *.csv in results/, outputs/, logs/
3. Ground truth paths:    look in eval scripts for data loading (dataset paths, GT references)
4. Experiment tracker:    EXPERIMENT_TRACKER.md, EXPERIMENT_LOG.md
5. Paper claims:          NARRATIVE_REPORT.md, paper/sections/*.tex, PAPER_PLAN.md
6. Config files:          *.yaml, *.toml, *.json configs with metric definitions

DO NOT summarize, interpret, or explain any file content. Only collect paths.

Step 2: Send to Reviewer (GPT-6-Astra via Codex MCP)

Pass ONLY file paths and the audit checklist to the reviewer. The reviewer reads everything directly.

Resolve this reviewer slot first; the following is the backend=codex example.

spawn_agent:
  model: gpt-6-astra
  reasoning_effort: ultra
  message: |
    You are an experiment integrity auditor. Start from the assumption that the
    evaluation is compromised somewhere — your job is to find where. Be
    adversarial. Trust nothing the author tells you — verify everything
    yourself. Read ALL files listed below and check for the following fraud
    patterns.

    Files to read:
    - Evaluation scripts: [list paths]
    - Result files: [list paths]
    - Experiment tracker: [list paths]
    - Paper claims: [list paths]
    - Config files: [list paths]

    ## Audit Checklist

    ### A. Ground Truth Provenance
    For each evaluation script:
    1. Where does "ground truth" / "reference" / "target" come from?
    2. Is it loaded from the DATASET, or generated/derived from MODEL OUTPUTS?
    3. If derived: is it explicitly labeled as proxy evaluation?
    4. Are official eval scripts used when available for this benchmark?
    FAIL if: GT is derived from model outputs without explicit proxy labeling.

    ### B. Score Normalization
    For each metric computation:
    1. Is any metric divided by max/min/mean of the model's OWN output?
    2. Are raw scores reported alongside any normalized scores?
    3. Are any scores suspiciously close to 1.0 or 100%?
    FAIL if: Normalization denominator comes from prediction statistics.

    ### C. Result File Existence
    For each claim in the paper/narrative:
    1. Does the referenced result file actually exist?
    2. Does the claimed metric key exist in that file?
    3. Does the claimed NUMBER match what's in the file?
    4. Is the experiment tracker status DONE (not TODO/IN_PROGRESS)?
    FAIL if: Claimed results reference nonexistent files or mismatched numbers.

    ### D. Dead Code Detection
    For each metric function defined in eval scripts:
    1. Is it actually CALLED in any evaluation pipeline?
    2. Does its output appear in any result file?
    WARN if: Metric functions exist but are never called.

    ### E. Scope Assessment
    1. How many scenes/datasets/configurations were actually tested?
    2. How many seeds/runs per configuration?
    3. Does the paper use words like "comprehensive", "extensive", "robust"?
    4. Is the actual scope sufficient for those claims?
    WARN if: Scope language exceeds actual evidence.

    ### F. Evaluation Type Classification
    Classify each evaluation as:
    - real_gt: uses dataset-provided ground truth
    - synthetic_proxy: uses model-generated reference
    - self_supervised_proxy: no GT by design
    - simulation_only: simulated environment
    - human_eval: human judges

    ## Output Format

    For each check (A-F), report:
    - Status: PASS | WARN | FAIL
    - Evidence: exact file:line references
    - Details: what specifically was found

    Overall verdict: PASS | WARN | FAIL
    
    Be thorough. Read every eval script line by line.

Step 3: Parse and Write Report (Executor — Codex)

Parse the reviewer's response and write EXPERIMENT_AUDIT.md:

# Experiment Audit Report

**Date**: [today]
**Auditor**: [actual selected model and effort; fresh read-only session;
actual family comparison and completion status]
**Project**: [project name]

## Overall Verdict: [PASS | WARN | FAIL]

## Integrity Status: [pass | warn | fail]

## Checks

### A. Ground Truth Provenance: [PASS|WARN|FAIL]
[details + file:line evidence]

### B. Score Normalization: [PASS|WARN|FAIL]
[details]

### C. Result File Existence: [PASS|WARN|FAIL]
[details]

### D. Dead Code Detection: [PASS|WARN|FAIL]
[details]

### E. Scope Assessment: [PASS|WARN|FAIL]
[details]

### F. Evaluation Type: [real_gt | synthetic_proxy | ...]
[classification + evidence]

## Action Items
- [specific fixes if WARN or FAIL]

## Claim Impact
- Claim 1: [supported | needs qualifier | unsupported]
- Claim 2: ...

Also write EXPERIMENT_AUDIT.json for machine consumption:

{
  "audit_skill": "experiment-audit",
  "verdict": "WARN",
  "reason_code": "scope_exceeds_evidence",
  "summary": "Two-scene evaluation supports a qualified claim only.",
  "audited_input_hashes": {"results/eval.json": "sha256:<hash>"},
  "trace_path": ".aris/traces/experiment-audit/2026-04-10_run01/",
  "agent_id": "agent_019f...",
  "verdict_id": "agent_019f...",
  "executor_model": "codex-gpt-6-astra",
  "executor_family": "openai",
  "reviewer_model": "gpt-6-astra",
  "reviewer_family": "openai",
  "reviewer_reasoning": "ultra",
  "review_independence": "<actual family comparison>",
  "acceptance_status": "<actual acceptance status>",
  "generated_at": "2026-04-10T00:00:00Z",
  "date": "2026-04-10",
  "auditor": "gpt-6-astra-ultra",
  "overall_verdict": "warn",
  "integrity_status": "warn",
  "checks": {
    "gt_provenance": {"status": "pass", "details": "..."},
    "score_normalization": {"status": "warn", "details": "..."},
    "result_existence": {"status": "pass", "details": "..."},
    "dead_code": {"status": "pass", "details": "..."},
    "scope": {"status": "warn", "details": "..."},
    "eval_type": "real_gt"
  },
  "claims": [
    {"id": "C1", "impact": "supported"},
    {"id": "C2", "impact": "needs_qualifier"}
  ]
}

Step 4: Print Summary

🔬 Experiment Audit Complete

  GT Provenance:      ✅ PASS — real dataset GT used
  Score Normalization: ⚠️ WARN — boundary metric uses self-reference
  Result Existence:    ✅ PASS — all files exist, numbers match
  Dead Code:           ✅ PASS — all metric functions called
  Scope:               ⚠️ WARN — 2 scenes, paper says "comprehensive"

  Overall: ⚠️ WARN
  
  See EXPERIMENT_AUDIT.md for details.

Integration with Other Skills

Automatic in /research-pipeline (advisory, never blocks)

When integrated into the pipeline, this skill runs automatically after /experiment-bridge and before /auto-review-loop:

/experiment-bridge → results ready
    ↓
/experiment-audit (automatic, advisory)
    ├── PASS  → continue normally
    ├── WARN  → print ⚠️ warning, continue, tag claims as [INTEGRITY: WARN]
    └── FAIL  → print 🔴 alert, continue, tag claims as [INTEGRITY CONCERN]
    ↓
/auto-review-loop → proceeds with integrity tags visible to reviewer

Never blocks the pipeline. Even on FAIL, the pipeline continues — but claims carry visible integrity tags.

Read by /result-to-claim (if exists)

if EXPERIMENT_AUDIT.json exists:
    read integrity_status
    attach to verdict: {claim_supported: "yes", integrity_status: "warn"}
    if integrity_status == "fail":
        downgrade verdict display: "yes [INTEGRITY CONCERN]"
else:
    verdict as normal, integrity_status = "unavailable"
    mark as "provisional — no integrity audit"

Read by /paper-write (if exists)

if EXPERIMENT_AUDIT.json exists AND integrity_status == "fail":
    add footnote to affected claims: "Note: integrity audit flagged concerns with this evaluation"

Key Rules

• Reviewer independence: executor collects paths, reviewer judges. Period.
• Never block: warn loudly, never halt the pipeline.
• File-as-switch: no EXPERIMENT_AUDIT.md = skill was never run = zero impact on existing behavior.
• Review class: base Codex is same-family provisional; a completed positive selected review may claim cross-family accepted only with known different model families.
• Honest about limits: the audit catches common patterns, not all possible fraud. It is a safety net, not a guarantee.

Acknowledgements

Motivated by community-reported integrity issues (#57, #131) where executor agents created fake ground truth and self-normalized scores.

Review Tracing

After each reviewer agent call, save the trace following shared-references/review-tracing.md (Policy C — forensic; never silently skip). Use save_trace.sh (resolved per the chain in shared-references/integration-contract.md §2) or write files directly to .aris/traces/<skill>/<date>_run<NN>/. Respect the --- trace: parameter (default: full).