- name
SKILL.md
스킬 파일은 게시자의 저장소에 그대로 있습니다.
복사한 명령을 터미널에 붙여넣습니다. 이 SKILL.md가 해당 에이전트에 설치됩니다.
npx skills add omagents/omagents --skill hyperplan -a cursor -y복사한 명령을 터미널에 붙여넣습니다. 이 SKILL.md가 해당 에이전트에 설치됩니다.
npx skills add omagents/omagents --skill hyperplan -a claude-code -y복사한 명령을 터미널에 붙여넣습니다. 이 SKILL.md가 해당 에이전트에 설치됩니다.
npx skills add omagents/omagents --skill hyperplan -a codex -y복사한 명령을 터미널에 붙여넣습니다. 이 SKILL.md가 해당 에이전트에 설치됩니다.
npx skills add omagents/omagents --skill hyperplan -a grok -y복사한 메시지를 Grok Bot 채팅에 보냅니다. Bot이 GitHub에서 이 SKILL.md를 가져와 모든 Bot이 공유하는 스킬에 추가합니다.
Install this skill: https://github.com/omagents/omagents/tree/HEAD/packages/core/skills/hyperplan게시자 제공
Dispatch 3 parallel critic subagents, each attacking a plan from a different perspective. The loop engine tracks completion; results are synthesized into a prioritized report.
loop_engine.py init hyperplan '[
{"critic": "security", "description": "Security critic: vulnerabilities, input validation, auth bypass, data exposure"},
{"critic": "architecture", "description": "Architecture critic: coupling, scalability, design patterns, maintainability"},
{"critic": "edge-cases", "description": "Edge cases critic: unhandled scenarios, error paths, race conditions, empty states"}
]'
Launch all 3 critics as background subagents in a single response:
task(background=true, description="Security critic", prompt="
You are a security critic. Review the following plan and find security issues.
Plan:
<insert plan here>
Analyze for:
1. Input validation gaps
2. Authentication/authorization bypass
3. Data exposure (logs, errors, responses)
4. Injection vulnerabilities
5. Dependency risks
Output: List of findings with severity (CRITICAL/HIGH/MEDIUM/LOW) and specific file/line references.
")
task(background=true, description="Architecture critic", prompt="
You are an architecture critic. Review the following plan and find design issues.
Plan:
<insert plan here>
Analyze for:
1. Tight coupling between modules
2. Scalability bottlenecks
3. Missing abstractions or over-engineering
4. Error handling strategy
5. Testability concerns
Output: List of findings with severity (CRITICAL/HIGH/MEDIUM/LOW) and specific design recommendations.
")
task(background=true, description="Edge cases critic", prompt="
You are an edge cases critic. Review the following plan and find unhandled scenarios.
Plan:
<insert plan here>
Analyze for:
1. Empty/null/undefined inputs
2. Concurrent access / race conditions
3. Network failures and timeouts
4. Resource exhaustion (memory, disk, connections)
5. Boundary conditions (0, max, negative, unicode)
Output: List of findings with severity (CRITICAL/HIGH/MEDIUM/LOW) and specific scenarios.
")
As each background task completes, record the result:
loop_engine.py complete hyperplan <id> "Found 3 issues: 1 CRITICAL, 2 MEDIUM"
If a critic fails to produce results:
loop_engine.py fail hyperplan <id> "Critic timed out"
When all critics are done:
loop_engine.py summary hyperplan
Generate a consolidated report:
## Hyperplan Review Report
### Summary
- Security: 1 CRITICAL, 2 MEDIUM
- Architecture: 0 CRITICAL, 3 MEDIUM
- Edge Cases: 0 CRITICAL, 1 HIGH, 2 LOW
### Critical Issues (must fix before implementation)
1. [SECURITY] SQL injection in search endpoint (src/api/search.ts:42)
Fix: Use parameterized queries
### High Priority (should fix before implementation)
2. [EDGE] No timeout on external API calls (src/services/external.ts:15)
Fix: Add 30s timeout with retry
### Medium Priority (address during implementation)
3. [ARCH] Search module directly imports from auth module
Fix: Introduce interface in shared/types.ts
4. [SECURITY] Error messages leak internal paths
Fix: Sanitize error responses
5. [ARCH] No retry strategy for failed jobs
Fix: Add exponential backoff in queue processor
6. [EDGE] Empty search query returns all records
Fix: Return empty result or validation error
### Low Priority (nice to have)
7. [EDGE] Unicode characters in usernames not handled
Fix: Add unicode normalization
8. [EDGE] Negative page numbers cause 500 error
Fix: Validate and return 400
Adversarial plan review with 3 parallel critics using loop engine tracking. Each critic attacks the plan from a different angle: security, architecture, and edge cases. Trigger when the user says 'hyp
---name: hyperplandescription: Adversarial plan review with 3 parallel critics using loop engine tracking. Each critic attacks the plan from…---# packages/core/skills/hyperplan/SKILL.md개요
Adversarial plan review with 3 parallel critics using loop engine tracking. Each critic attacks the plan from a different angle: security, architecture, and edge cases. Trigger when the user says 'hyperplan', 'review plan', 'critique plan', or wants adversarial analysis of a design.
FAQ
게시자에게서 packages/core/skills/hyperplan/SKILL.md를 다운로드하거나 경로를 복사해 저장소를 엽니다. PlugMCP는 스킬을 목록에 올릴 뿐 파일을 호스팅하지 않습니다.
게시자가 https://github.com/omagents/omagents에서 관리합니다.
Claude, Cursor, Codex, Grok Build는 게시자의 저장소에서 추가한 SKILL.md를 읽을 수 있습니다. Grok Bot은 채팅에서 스킬의 GitHub 폴더 링크를 보내면 설치합니다.