Repair plan

Prioritized gold repairs

The Repair Planner ranks each fix with impact, confidence, effort, and risk so the demo can separate quick wins from approval-gated changes.

Monthly savings

$1,670

Projected after repair execution

Annual savings

$20,040

Business value for CTO and finance teams

Carbon reduction

92 kgCO2e

Estimated monthly reduction

Engineering time

44h/mo

Recovered from CI and duplicate workflow waste

Repair Priority Score = impact x confidence / effort / risk

Impact combines monthly cost, carbon, and severity. Confidence comes from agent evidence. Effort and risk reduce the score so KintsugiGuard AI does not blindly chase risky savings.

Repair Lanes

Execution grouping by autonomy and risk.

Before and After

Estimated impact after gold repairs are applied.

Quick wins

Low-risk, high-confidence fixes the agent can prepare first.

Redundant CI/CD jobs repeat the same work

5.58
HighLowEffort Low

Share dependency cache, split changed-package checks, and collapse duplicate transpilation steps.

Owner
Developer experience owner
Verification
Run workflow on three representative PRs and compare runtime plus failure rate.

Bloated Docker image

4.49
HighLowEffort Medium

Use a multi-stage build, prune package caches, and ship only production artifacts.

Owner
Service team owner
Verification
Run targeted regression tests and compare resource metrics.

Medium-term fixes

Worth scheduling after the fast repairs land.

Idle scheduled workloads stay warm all day

3.36
HighLowEffort Medium

Convert daily batch workers to event-driven jobs with a shutdown policy outside processing windows.

Owner
DevOps platform owner
Verification
Canary deploy with utilization, latency, and rollback guardrails.

Unused dependencies in runtime package

3.12
MediumLowEffort Low

Remove unused packages and consolidate overlapping date, schema, and HTTP client libraries.

Owner
Service team owner
Verification
Run targeted regression tests and compare resource metrics.

Excessive duplicate AI model calls

2.06
HighMediumEffort Medium

Add prompt-hash caching, retry de-duplication, and lower-cost routing for deterministic summaries.

Owner
AI platform owner
Verification
Compare quality sample set before and after cache or routing policy.

Duplicate API endpoint and transform function

1.3
MediumMediumEffort Low

Merge duplicate reporting endpoints behind a shared transform with compatibility tests.

Owner
Service team owner
Verification
Run targeted regression tests and compare resource metrics.

Missing cache for high-volume API reads

0.93
MediumMediumEffort Medium

Introduce a short-lived cache for repeated tenant dashboard reads and invalidate on writes.

Owner
Service team owner
Verification
Run targeted regression tests and compare resource metrics.

Human approval

High-impact or higher-risk changes that need an owner before execution.

Over-provisioned cloud service tier

3.45
CriticalMediumEffort Medium

Right-size ECS tasks to the observed p95 CPU and memory profile, then apply autoscaling guardrails.

Owner
DevOps platform owner
Verification
Canary deploy with utilization, latency, and rollback guardrails.

Suggested pull request summary

Slim runtime image and production dependencies. Add CI dependency cache and selective workflow execution. Introduce budget-aware AI API caching and routing. Apply cloud right-sizing behind canary and approval gates. Attach before/after cost, carbon, and runtime checks to the repair ledger.

Next screen

Agentic payment moment

The X402 Payment Agent buys a specialist carbon audit result only after a budget policy approves the price.

Simulate X402 payment