Help us improve
Share bugs, ideas, or general feedback.
From space-agents
Guides interactive code reviews through conversation, spawning agents for quality, security, performance, and simplification. Categorizes issues by priority and creates Beads.
npx claudepluginhub thebrownproject/space-agents --plugin space-agentsHow this skill is triggered — by the user, by Claude, or both
Slash command
/space-agents:exploration-reviewThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Review code through conversation. This is collaborative analysis, not a report dump. You guide the review, ask questions, spawn specialized agents, and work through findings together.
Performs exhaustive code reviews using multi-agent analysis and ultra-thinking, with Git worktree isolation and knowledge recall from bead memory.
Conducts code reviews of pull requests and changes for correctness, security, performance, and maintainability with actionable feedback grouped by severity levels.
Analyzes code quality via cleanup analysis (dead code, duplication, complexity) and multi-perspective reviews (maintainer, architect, user advocate, security, performance).
Share bugs, ideas, or general feedback.
Review code through conversation. This is collaborative analysis, not a report dump. You guide the review, ask questions, spawn specialized agents, and work through findings together.
Ask what to review:
Use AskUserQuestion:
"Which areas should I focus on?"
Options:
- Quality (readability, structure, patterns)
- Security (secrets, injection, validation)
- Performance (algorithms, queries, optimization)
- Simplification (dead code, over-engineering, DRY)
- All of the above
Based on selection, spawn with run_in_background: true:
| Category | Agent | Focus |
|---|---|---|
| Quality | space-agents:review-quality | Readability, naming, complexity, patterns |
| Security | space-agents:review-security | Secrets, injection, auth, OWASP |
| Performance | space-agents:review-performance | Algorithms, queries, caching, bundle |
| Simplification | space-agents:review-code-simplifier | Dead code, over-engineering, DRY, bloat |
Continue conversation while agents work. Check results with TaskOutput block: false.
Categorize all findings:
| Priority | Meaning | Action |
|---|---|---|
| Critical | Security vulnerability, data loss risk, broken functionality | Must fix before merge |
| Warning | Code smell, maintainability issue, potential bug | Should fix |
| Suggestion | Style improvement, optimization opportunity | Consider improving |
For each finding:
Red flags to watch for:
Spawn with run_in_background: true, continue conversation immediately:
space-agents:review-quality - Code quality and maintainabilityspace-agents:review-security - Security vulnerabilities and risksspace-agents:review-performance - Performance issues and optimizationsspace-agents:review-code-simplifier - Dead code, over-engineering, DRY violationsAlways use AskUserQuestion for every question in review. Prefer multiple choice when you can anticipate likely answers.
When review is complete:
Present findings organized by priority:
## Review Summary
### Critical (must fix)
- [Issue with file:line reference]
### Warnings (should fix)
- [Issue with file:line reference]
### Suggestions (consider)
- [Issue with file:line reference]
### What's Good
- [Positive observations]
Ask user:
AskUserQuestion:
"Want to create Beads to track these issues?"
Options:
- "Yes, create bugs for Critical/Warning" - Track issues that need fixing
- "Yes, create tasks for all" - Track everything including suggestions
- "No, I'll handle it" - Skip Bead creation
If creating Beads:
bd create --type=bug for Critical/Warning issuesbd create --type=task for Suggestions