etanhey/review-router
skills/golem-powers/review-router/SKILL.md
Dynamic code review routing with automatic fallback chain when primary reviewer is unavailable. Routes to CodeRabbit, Macroscope, requesting-code-review, or Cursor CLI. Triggers on: 'review-router', 'route review', 'reviewer unavailable'. NOT for: general code review workflow (use /code-review), receiving review feedback (use /superpowers:receiving-code-review).
$ install --global
skillsauthnpx skillsauth add etanhey/golems review-routerInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: May 1, 2026, 3:47 AM349.8s2 files scanned
SKILL.md
- name:
- review-router
- description:
- Dynamic code review routing with automatic fallback chain when primary reviewer is unavailable. Routes to CodeRabbit, Macroscope, requesting-code-review, or Cursor CLI. Triggers on: 'review-router', 'route review', 'reviewer unavailable'. NOT for: general code review workflow (use /code-review), receiving review feedback (use /superpowers:receiving-code-review).
Review Router
Routes reviews to available tools with automatic fallback. The overnight sprint had CodeRabbit rate-limited and Greptile expired on 4/5 golems PRs. Workers fell back randomly. This skill defines the deterministic fallback chain.
Note: Greptile was evaluated but excluded — expired tokens are not auto-recoverable overnight.
Fallback Chain (in order)
-
CodeRabbit (
coderabbit:code-reviewersubagent orcr review --plain)- Check:
cr --versionsucceeds AND no rate limit in last 5 min - Timeout: 120 seconds
- On failure: log "CodeRabbit unavailable: {reason}", fall through
- Check:
-
Macroscope (GitHub App — posts as PR comment)
- Check:
gh api repos/{owner}/{repo}/installationsincludes Macroscope - Timeout: 300 seconds (it's slow)
- On failure: log "Macroscope unavailable", fall through
- Check:
-
superpowers:requesting-code-review (Claude subagent)
- Check: always available (it's a subagent)
- Cost: ~50K tokens per review
- On failure: should not fail, but if it does, fall through
-
Cursor CLI — Red/Blue Profiles (external tool — read-only audit mode)
- Check:
cursor --versionsucceeds <cursor-max-mode-model>below: substitute Cursor's current Max Mode model ID (verify viacursor agent --helpor the Cursor changelog — IDs change between versions; the user's preference is Cursor on Auto, but Red/Blue audits historically pin to Max Mode for harsher critique).- Red Team (security/reliability):
cursor agent --output-format text --model "<cursor-max-mode-model>" "$(cat ~/.claude/commands/code-review/references/red-team-prompt.md | sed 's/{{REPO_CONTEXT}}/...context.../') DIFF: $(gh pr diff)" - Blue Team (quality/architecture):
cursor agent --output-format text --model "<cursor-max-mode-model>" "$(cat ~/.claude/commands/code-review/references/blue-team-prompt.md | sed 's/{{REPO_CONTEXT}}/...context.../') DIFF: $(gh pr diff)" - Timeout: 180 seconds per profile
- Run BOTH for comprehensive coverage (replaces single generic prompt)
- On failure: WARN — all reviewers exhausted, post to collab and wait
- See
/code-reviewworkflows:workflows/red-team.mdandworkflows/blue-team.md
- Check:
All Reviewers Exhausted
If all 4 tiers fail:
- Log each failure reason
- Post a BLOCKED message in the collab file
- Do NOT merge without review — wait for human or tool recovery
Minimum Review Requirement
2 independent reviews before merge. Preferred: 1 red team + 1 blue team. If only 1 reviewer tool is available, run it with both profiles:
- Pass 1: Red team prompt (security, reliability, crash scenarios)
- Pass 2: Blue team prompt (architecture, quality, test coverage)
Deduplicate findings: red H findings take priority over blue suggestions on the same code.
H/M/L Classification (from overnight retro)
After collecting review findings, classify each:
- HIGH/CRITICAL → Fix immediately, block merge
- MEDIUM → Fix before merge if <15 min effort, else document as follow-up
- LOW → Merge, create follow-up issue if warranted
Usage
# In /pr-loop, after PR creation:
/review-router
# Standalone:
/review-router --pr 42 --repo golems
Related Skills
etanhey/phoenix-human-view
tools
VerifiedTrustedCommunity
The human-eval UX contract for Phoenix views: turn-by-turn scrollable replay (not a scorecard), hide-but-copyable IDs, collapsed thinking, identity chips, tool filters, tiny frozen starter datasets, mark-wrong-in-thread, mobile-first. Use when: building or reviewing ANY Phoenix/eval view, annotation UI, session replay, or human-grading surface. Triggers: phoenix view, eval UI, annotation view, session replay, human eval UX, grading interface. NOT for: Phoenix data pipelines/ingest (capture scripts have their own specs).
3SKILL.mdUpdated Jun 7, 2026
etanhey/mac-systems
tools
VerifiedTrustedCommunity
macOS systems specialist — AppKit NSPanel architecture, launchd services, socket activation, MCP bridge resilience, syspolicyd, and high-frequency SwiftUI dashboards. Use when building menu-bar apps, LaunchAgents, debugging syspolicyd/Gatekeeper/TCC, resilient UDS/MCP bridges, or SwiftUI dashboards at 10Hz+.
3SKILL.mdUpdated Jun 7, 2026
etanhey/judge-fleet
development
VerifiedTrustedCommunity
Bulk LLM-judging protocol for fleet-dispatched verdict runs (KG cluster, eval harness). Use when: dispatching or running judge workers (J1/J2/RT), planning bulk-apply from verdict JSONL, or triaging evidence_degraded outputs. Triggers: judge fleet, bulk judge, R3 verdicts, kg-judge, RT gate, evidence_degraded. NOT for: single-item code review, Phoenix view UX (use phoenix-human-view), or non-judge eval pipelines.
3SKILL.mdUpdated Jun 7, 2026
etanhey/fleet-wrap
development
VerifiedTrustedCommunity
Quiet-down protocol for sprint close: when the fleet wraps, delete ALL polling crons and monitors, send ONE final dashboard + ONE message, then go SILENT. Use when: fleet wraps, all workers done, overnight queue exhausted, sprint close, Etan asleep/away with nothing approved left. Triggers: fleet wrap, wrap the fleet, stand down, going quiet, sprint close. NOT for: mid-sprint monitoring (keep your loops), spawning a successor (use /session-handoff first).
3SKILL.mdUpdated Jun 7, 2026