Aimi Review

Perform code reviews using parallel aimi-native review agents with findings synthesis.

Step 1: Determine Review Target

<review_target> #$ARGUMENTS </review_target>

Detect Target Type

1. PR number (numeric): Fetch PR with gh pr view $ARGUMENTS --json title,body,files,headRefName,baseRefName
2. GitHub URL: Extract PR number, then fetch as above
3. Branch name: Compare against main/master with git diff main...$ARGUMENTS --name-only
4. Empty (no arguments): Review current branch against main/master

Setup

# Get changed files
gh pr view [number] --json files --jq '.files[].path'
# OR for branch comparison:
git diff main...HEAD --name-only

Read the changed files to understand the PR content. Collect the diff for agent context.

Protected Artifacts

These paths must never be flagged for deletion or removal by any review agent:

• .aimi/plans/*.md — Plan files
• .aimi/solutions/*.md — Solution documents
• .aimi/tasks/*.json — Task files
• .aimi/brainstorms/*.md — Brainstorm documents

Step 2: Run Default Review Agents (Parallel)

Run these agents in parallel using the Task tool:

Task subagent_type="aimi-engineering:review:aimi-architecture-strategist"
  prompt: "Review this code for architectural compliance:
           [PR content / diff summary]"

Task subagent_type="aimi-engineering:review:aimi-security-sentinel"
  prompt: "Perform security audit on this code:
           [PR content / diff summary]"

Task subagent_type="aimi-engineering:review:aimi-code-simplicity-reviewer"
  prompt: "Review this code for simplicity and minimalism:
           [PR content / diff summary]"

Task subagent_type="aimi-engineering:review:aimi-performance-oracle"
  prompt: "Analyze this code for performance issues:
           [PR content / diff summary]"

Task subagent_type="aimi-engineering:review:aimi-agent-native-reviewer"
  prompt: "Verify new features are agent-accessible:
           [PR content / diff summary]"

Task subagent_type="aimi-engineering:research:aimi-learnings-researcher"
  prompt: "Search .aimi/solutions/ for past issues related to this PR:
           [PR content / diff summary]"

Step 3: Run Conditional Agents (If Applicable)

Migration Agents

Run ONLY when PR contains database migrations, schema changes, or data backfills.

Detection: Check if changed files include db/migrate/*.rb, db/schema.rb, migration scripts, or data backfill tasks.

Task subagent_type="aimi-engineering:review:aimi-schema-drift-detector"
  prompt: "Detect unrelated schema.rb changes:
           [PR content / diff summary]"

Task subagent_type="aimi-engineering:review:aimi-data-migration-expert"
  prompt: "Validate ID mappings and migration safety:
           [PR content / diff summary]"

Task subagent_type="aimi-engineering:review:aimi-deployment-verification-agent"
  prompt: "Create deployment checklist with verification queries:
           [PR content / diff summary]"

Language-Specific Agents

Detect primary language from changed files and run the appropriate reviewer:

File Extensions	Agent
*.rb, Gemfile, *.erb	aimi-engineering:review:aimi-kieran-rails-reviewer
*.ts, *.tsx, *.js	aimi-engineering:review:aimi-kieran-typescript-reviewer
*.py	aimi-engineering:review:aimi-kieran-python-reviewer

For Rails projects, also consider running:

• aimi-engineering:review:aimi-dhh-rails-reviewer for Rails convention checks
• aimi-engineering:review:aimi-julik-frontend-races-reviewer for Stimulus/JS race conditions

Step 4: Findings Synthesis

Consolidate Results

1. Collect findings from all parallel agents
2. Surface learnings-researcher results: flag relevant past solutions as "Known Pattern" with links to .aimi/solutions/ files
3. Discard any findings that recommend deleting protected artifacts (see Step 1)
4. Remove duplicate or overlapping findings

Categorize by Severity

• P1 CRITICAL — Blocks merge: security vulnerabilities, data corruption risks, breaking changes
• P2 IMPORTANT — Should fix: performance issues, architectural concerns, code quality problems
• P3 NICE-TO-HAVE — Enhancements: minor improvements, cleanup, documentation

Estimate Effort

For each finding: Small (< 30 min), Medium (30 min - 2 hours), Large (> 2 hours)

Step 5: Aimi-Branded Report

## Review Complete

**Review Target:** [PR title or branch name]
**Branch:** [branch-name]

### Findings Summary

- **Total Findings:** [X]
- **P1 CRITICAL:** [count] - BLOCKS MERGE
- **P2 IMPORTANT:** [count] - Should Fix
- **P3 NICE-TO-HAVE:** [count] - Enhancements

### P1 - Critical (Must Fix Before Merge)

1. **[Finding title]** — [description]
   - Location: [file:line]
   - Impact: [what breaks]
   - Fix: [how to fix]
   - Effort: [Small/Medium/Large]

### P2 - Important (Should Fix)

1. **[Finding title]** — [description]
   - Location: [file:line]
   - Recommendation: [how to improve]
   - Effort: [Small/Medium/Large]

### P3 - Nice-to-Have

1. **[Finding title]** — [description]
   - Suggestion: [improvement]

### Review Agents Used

- aimi-architecture-strategist
- aimi-security-sentinel
- aimi-code-simplicity-reviewer
- aimi-performance-oracle
- aimi-agent-native-reviewer
- aimi-learnings-researcher
- [conditional agents if run]

### Next Steps

1. **Address P1 findings** — Critical issues must be fixed before merge
2. **Run `/aimi:execute`** — Continue autonomous execution
3. **Run `/aimi:status`** — Check current task progress

Error Handling

Failure	Action
No review target found	Ask user to specify PR number or branch
Agent fails	Proceed with available results, note in report
No changed files	Report "No changes to review"
gh CLI not installed	Fall back to git diff for branch comparison