Assurance overview

Spec set: agent education system, data pipeline, internal tool cli · 5 metrics × 4 conditions · A concise view of assurance posture for governance and rollout decisions

antigravity

claude_code

cursor_agent

replit_agent

complexity_mean

2.39 cc

Best observed posture: replit_agent
Most challenging posture: claude_code (3.35)

Average structural complexity per function, indicating how maintainable the code is likely to be.

correction_freq

0.00 per_kkey

Best observed posture: antigravity
Most challenging posture: replit_agent (0.00)

Frequency of corrective edits during the session, indicating how much rework the workflow required.

duplication_pct

0.00 %

Best observed posture: claude_code
Most challenging posture: replit_agent (9.56)

Share of code that appears in repeated 6-line patterns, signalling maintainability debt.

hallucinations

0.00 count

Best observed posture: claude_code
Most challenging posture: replit_agent (1.33)

Features implemented outside the approved specification, creating delivery and compliance risk.

security_density

0.00 per_kloc

Best observed posture: replit_agent
Most challenging posture: claude_code (9.65)

Security findings per 1,000 lines of Python, expressed as a governance-relevant density signal.

Assurance view

Use the toggle to move between calibrated assurance scores and raw values. Calibrated scores are framed around practical decision thresholds; raw values preserve the underlying measurement scale.

This view is designed to support adoption decisions and governance conversations, not to replace expert judgment.

Adoption ranking

A directional view of relative assurance posture. Illustrative only — this is not a weighted procurement score and should be read alongside the metric guidance below.

antigravity

avg rank 2.2

🏆 1

claude_code

avg rank 2.2

🏆 3

cursor_agent

avg rank 2.2

🏆 1

replit_agent

avg rank 2.2

🏆 3

Assurance heatmap

Green indicates stronger assurance posture; red highlights areas that warrant attention.

Metric assurance review

Select a metric to review the signal, the threshold guidance, and the implication for rollout.

complexity_mean

correction_freq

duplication_pct

hallucinations

security_density

complexity_mean

Average structural complexity per function, indicating how maintainable the code is likely to be.

Unit: cc · Lower is better

Decision guidance

Decision guidance: lower is better. Values at or below 3 are broadly sustainable; 3 to 6 signals rising maintainability risk; above 6 is likely to become brittle in production.

Adoption implication

Adoption implication: code that is structurally complex is harder to maintain, review, and govern at scale.

Condition	Value	Rank
antigravity	2.601	2
claude_code	3.354	4
cursor_agent	2.723	3
replit_agent	2.388	1

correction_freq

Frequency of corrective edits during the session, indicating how much rework the workflow required.

Unit: per_kkey · Lower is better

Decision guidance

Decision guidance: lower is better. Values at or below 10 are efficient; 10 to 25 indicates repeated editing effort; above 25 suggests a poor interaction loop for real-world use.

Adoption implication

Adoption implication: high correction frequency points to friction that can erode developer trust and slow delivery.

Condition	Value	Rank
antigravity	0.000	1
claude_code	0.000	1
cursor_agent	0.000	1
replit_agent	0.000	1

duplication_pct

Share of code that appears in repeated 6-line patterns, signalling maintainability debt.

Unit: % · Lower is better

Decision guidance

Decision guidance: lower is better. Values at or below 5 are healthy; 5 to 10 suggests avoidable copy-and-paste debt; above 10 is a strong sign of maintainability problems.

Adoption implication

Adoption implication: high duplication increases the chance of inconsistent fixes and makes long-term stewardship harder.

Condition	Value	Rank
antigravity	4.258	3
claude_code	0.000	1
cursor_agent	0.904	2
replit_agent	9.556	4

hallucinations

Features implemented outside the approved specification, creating delivery and compliance risk.

Unit: count · Lower is better

Decision guidance

Decision guidance: lower is better. Zero is ideal; 1 to 3 indicates scope drift and trust risk; above 3 is a serious control failure.

Adoption implication

Adoption implication: a tool that ships features outside the spec creates procedural and compliance risk, even when it appears productive.

Condition	Value	Rank
antigravity	0.333	3
claude_code	0.000	1
cursor_agent	0.167	2
replit_agent	1.333	4

security_density

Security findings per 1,000 lines of Python, expressed as a governance-relevant density signal.

Unit: per_kloc · Lower is better

Decision guidance

Decision guidance: lower is better. Values at or below 50 are manageable; 50 to 100 signals rising governance risk; above 100 is a clear concern.

Adoption implication

Adoption implication: tools that produce frequent security issues should not be rolled out broadly without remediation and review.

Condition	Value	Rank
antigravity	1.475	2
claude_code	9.646	4
cursor_agent	5.934	3
replit_agent	0.000	1

Provenance metadata

{
  "conditions": [
    "claude_code",
    "cursor_agent",
    "replit_agent",
    "antigravity"
  ],
  "design": "4 conditions \u00d7 3 specs \u00d7 10 reps = 120 runs, 600 metric rows",
  "deviations": "see docs/PROTOCOL_DEVIATIONS.md",
  "kind": "main_study",
  "live_conditions": [
    "claude_code",
    "cursor_agent"
  ],
  "replay_conditions": [
    "replit_agent",
    "antigravity"
  ],
  "reps_per_cell": 10,
  "spec_files": [
    "specs/agent_education_system.yaml",
    "specs/data_pipeline.yaml",
    "specs/internal_tool_cli.yaml"
  ],
  "timestamp": "2026-06-01T08:41:11.491653+00:00"
}