Codeguru Diagnostics

Amazon CodeGuru Diagnostics

When to use

Any Amazon CodeGuru investigation — reviewer recommendations, profiler, security detector, repository association, CI/CD integration, or suppression rules.

Investigation workflow

Step 1 — Collect and triage

aws codeguru-reviewer list-repository-associations --query 'RepositoryAssociationSummaries[*].{Name:Name,State:State,Arn:AssociationArn}'
version: "1.0.0"
last_updated: "2025-04-12"
aws codeguru-reviewer list-code-reviews --type PullRequest --query 'CodeReviewSummaries[*].{Name:Name,State:State}'
version: "1.0.0"
last_updated: "2025-04-12"
aws codeguru-profiler list-profiling-groups

Step 2 — Domain deep dive

aws codeguru-reviewer describe-repository-association --association-arn 
aws codeguru-reviewer list-recommendations --code-review-arn 
aws codeguru-profiler describe-profiling-group --profiling-group-name 
aws codeguru-profiler get-profile --profiling-group-name  --accept application/json

Step 3 — Detailed investigation

aws codeguru-reviewer list-code-reviews --type RepositoryAnalysis
aws codeguru-profiler list-findings-reports --profiling-group-name  --start-time  --end-time 
aws cloudtrail lookup-events --lookup-attributes AttributeKey=EventSource,AttributeValue=codeguru-reviewer.amazonaws.com --max-results 20

Read references/guardrails.md before concluding on any CodeGuru issue.

Gotchas: Amazon CodeGuru

• Repository association must be ASSOCIATED before reviews work.
• Reviewer analyzes pull requests automatically but full scans must be triggered.
• Profiler agent must be integrated into the application runtime.
• Security detector findings are separate from code quality recommendations.
• Suppressed recommendations are hidden but not deleted.
• Profiler data takes 5-15 minutes to appear after agent starts.

Anti-hallucination rules

1. Always cite association ARNs, review ARNs, or API responses as evidence.
2. Reviewer needs repository association. Never assume auto-discovery.
3. Profiler requires agent integration. Never suggest it works without agent.
4. Spend no more than 2 minutes on any single hypothesis. Pivot if inconclusive.

14 runbooks

| Category | IDs | Covers | |----------|-----|--------| | A — Reviewer | A1-A3 | Recommendations, code reviews, full scans | | B — Profiler | B1-B3 | Agent setup, profiling data, anomalies | | C — Security | C1-C2 | Security detector, findings | | D — Repository | D1-D2 | Association, provider config | | E — CI/CD | E1-E2 | Pipeline integration, automation | | F — Suppression | F1-F2 | Suppression rules, recommendation management |

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: aws-samples
• Source: aws-samples/sample-ai-agent-skills
• License: MIT-0

Install and usage instructions live in the source repository linked above.