메뉴
Stress-test solutions using the STRIKE framework. Systematically attack proposals to find weaknesses before deployment.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | adversarial-reasoning |
| description | Stress-test solutions using the STRIKE framework. Systematically attack proposals to find weaknesses before deployment. |
| metadata | {"clawdbot":{"requires":{"env":[]}}} |
| user-invocable | true |
Validate and stress-test solutions by systematically attacking them.
AR requires a proposed solution to attack. If no solution exists, use ToT or BoT first.
Identify all possible ways the solution could fail:
For each attack vector, design and execute tests:
Score each vulnerability found:
For each vulnerability:
Based on findings:
Document all findings for future reference.
## AR Validation: [Solution Name]
### Solution Under Test
[Brief description of the proposed solution]
### Attack Surface Analysis
#### Attack Vector 1: [name]
- **Attack**: [how to break it]
- **Result**: [what happened]
- **Severity**: [Critical/High/Medium/Low]
- **Mitigation**: [fix or accept]
#### Attack Vector 2: [name]
...
### Vulnerability Summary
| Vector | Severity | Exploitable | Mitigation |
|--------|----------|-------------|------------|
| [name] | High | Yes | [action] |
| [name] | Medium | No | Accept |
### Stress Test Results
- Load test: [result]
- Failure injection: [result]
- Edge cases: [result]
### Verdict
**[KILL / REVISE / KEEP]**
Confidence: X%
### Required Changes (if REVISE)
1. [change 1]
2. [change 2]
### Residual Risks (if KEEP)
- [risk 1]: [acceptance rationale]
Solution: JWT authentication for API
Attack Vectors Tested:
Verdict: REVISE
Then: KEEP with 85% confidence
Self-improving agent architecture using ChromaDB for continuous learning, self-evaluation, and improvement storage. Agents maintain separate memory collections for learned patterns, performance metrics, and self-assessments without modifying their static .md configuration.
Guide for manually creating custom skills from scratch with templates and validation. Use when designing NEW skills with specific requirements. For auto-generating skills from existing git repos, use /generate-skills instead.
Display QAVR (Q-Value Augmented Vector Retrieval) system status including mode, interaction counts, top memories by Q-value, and configuration.
Enables continuous self-improvement through learning from failures, user corrections, and capability gaps. Integrates with QAVR for learned memory ranking.
Exhaustive solution space exploration. Use when you need ALL viable options (not just the best), solution space is unknown, or you can't afford to miss alternatives.
Diagnostic reasoning for finding root causes. Generates multiple hypotheses, designs tests to eliminate each, and converges on the true cause through evidence.