Skip to main content

Stage 7: Review

Overview

Review validates that implementations meet acceptance criteria and quality standards before delivery. This stage covers code review, pull request creation, dashboard testing, prompt evaluation, and implementation validation against plans.

When You Enter This Stage

You enter Review after completing implementation work in Stage 6: Implementation.

NOTE

Prerequisites: Implementation complete with all changes committed. Use /clear to reset context before starting review.

Available Tools

Primary Agents

ToolTypeHow to InvokePurpose
task-reviewerAgentSelect task-reviewer agentReview implementation against the plan
pr-reviewAgentSelect pr-review agentEvaluate pull requests for quality
test-streamlit-dashboardAgentSelect test-streamlit-dashboard agentTest Streamlit dashboard implementations

Supporting Agents

ToolTypeHow to InvokePurpose
rpi-validatorAgentSelect rpi-validator agentValidate RPI workflow compliance
implementation-validatorAgentSelect implementation-validator agentCheck implementation against specifications
prompt-testerAgentSelect prompt-tester agentTest prompt engineering artifacts
prompt-evaluatorAgentSelect prompt-evaluator agentEvaluate prompt quality and effectiveness

Prompts and Instructions

ToolTypeHow to InvokePurpose
task-reviewPrompt/task-reviewStart a structured task review
pull-requestPrompt/pull-requestCreate a pull request for current changes
ado-create-pull-requestPrompt/ado-create-pull-requestCreate an ADO-linked pull request
doc-ops-updatePrompt/doc-ops-updateUpdate documentation alongside code changes
commit-messageInstructionAuto-activatedEnforces commit message conventions
community-interactionInstructionAuto-activatedEnforces community communication standards

Role-Specific Guidance

Engineers submit work for review and participate as peer reviewers. Tech Leads serve as primary reviewers, evaluating architecture alignment and code quality. Data Scientists review notebooks and dashboard outputs. Security Architects validate implementation against security requirements and compliance standards.

Starter Prompts

Implementation Review

Select task-reviewer agent:

Review today's changes to the authentication service against .copilot-tracking/plans/2025-01-15/auth-refactor-plan.instructions.md and check for missing input validation on the new endpoints
/task-review scope=today
/task-review plan=.copilot-tracking/plans/2025-01-15/pagination-plan.instructions.md changes=.copilot-tracking/changes/2025-01-15/pagination-changes.md research=.copilot-tracking/research/2025-01-15/pagination-research.md

Pull Request Workflow

/pull-request branch=origin/main excludeMarkdown=true
/ado-create-pull-request adoProject=hve-core baseBranch=origin/main isDraft=true workItemIds=54321,54322

Select pr-review agent:

Review the open PR for the payment processing refactor, focusing on breaking changes to the /api/payments endpoint and any exposed credentials in configuration files

Dashboard Testing

Select test-streamlit-dashboard agent:

Test the sensor monitoring dashboard at src/dashboards/sensor_monitor.py, verifying that temperature readings render within the 15-45°C expected range and all navigation links resolve correctly

Quality Validation

Select rpi-validator agent:

Validate phase 2 of .copilot-tracking/plans/2025-01-15/api-redesign-plan.instructions.md against .copilot-tracking/changes/2025-01-15/api-redesign-changes.md

Select implementation-validator agent:

Run full-quality validation on the files changed in src/services/auth/ against the architecture requirements in docs/architecture/auth-design.md

Select prompt-tester agent:

Execute .github/prompts/rpi/task-review.prompt.md literally in a sandbox to verify the review workflow produces expected validation outputs

Select prompt-evaluator agent:

Evaluate the execution log from .copilot-tracking/sandbox/2025-01-15-task-review-001/execution-log.md against the prompt quality criteria in .github/instructions/rpi/prompt-builder.instructions.md

Documentation Review

/doc-ops-update scope=docs/hve-guide/lifecycle validateOnly=true focus=accuracy

Stage Outputs and Next Stage

Review produces reviewed pull requests with feedback, validation reports, and approval decisions. Transition to Stage 8: Delivery when the PR is approved. Return to Stage 6: Implementation when rework is needed.

🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.