deployment-pipeline-design

Design multi-stage CI/CD pipelines with approval gates, security checks, and deployment orchestration. Use when architecting deployment workflows, setting up continuous delivery, or implementing GitOps practices.

1.12x

Quality

92%

Does it follow best practices?

Impact

91%

1.12x

Average score across 3 eval scenarios

Securityby

Passed

No known issues

Evaluation results

98%

17%

CI/CD Pipeline Design for a New Containerized Service

Pipeline stage structure and approval gates

Criteria

Without context

With context

Source stage present

100%

Build stage present

100%

Test stage present

100%

Security scan in test stage

80%

Staging deploy stage

100%

Integration tests after staging

71%

100%

Approval gate before production

100%

GitHub Actions environment for approval

100%

Production deploy stage

100%

Verification stage

71%

100%

Rollback stage or step

100%

Correct stage ordering

80%

100%

Fail fast: unit before E2E

90%

100%

Without context: $0.2093 · 53s · 8 turns · 9 in / 3,668 out tokens

With context: $0.5445 · 1m 45s · 25 turns · 655 in / 5,849 out tokens

87%

11%

Progressive Delivery for High-Stakes API Rollout

Progressive deployment strategies with rollback

Criteria

Without context

With context

Argo Rollouts kind

100%

Correct replica count

100%

Canary strategy used

100%

10% first canary step

100%

Progressive weights 25% and 50%

100%

Final 100% weight step

100%

5-minute pauses between steps

100%

30%

Health check retry loop

100%

curl -sf for health check

42%

100%

Automated rollback on failure

100%

Rollback uses kubectl rollout undo

60%

70%

Rollback targets correct resource

60%

70%

Without context: $0.2026 · 45s · 11 turns · 11 in / 2,611 out tokens

With context: $0.4718 · 1m 35s · 21 turns · 70 in / 5,260 out tokens

90%

Redesign a Slow, Incident-Prone Deployment Pipeline

Pipeline best practices and monitoring integration

Criteria

Without context

With context

Parallel independent jobs

100%

Dependency caching

100%

Artifact storage

100%

No hardcoded secrets

100%

Secret store reference

DORA metrics coverage

100%

Deployment Frequency metric

100%

Change Failure Rate metric

100%

MTTR metric

100%

Post-deploy error rate check

100%

1% error rate threshold

100%

Automated rollback mechanism

100%

Unit tests before E2E

75%

100%

Without context: $0.2947 · 1m 30s · 11 turns · 60 in / 5,396 out tokens

With context: $0.6515 · 2m 18s · 28 turns · 26 in / 7,586 out tokens

Repository: Dicklesworthstone/pi_agent_rust
Commit: d264bb8

Evaluated: 18 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

CI/CD Pipeline Design for a New Containerized Service Progressive Delivery for High-Stakes API Rollout Redesign a Slow, Incident-Prone Deployment Pipeline

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.