القائمة
Invoke the Devil's Advocate for steel-man + 10-dimension attack on any proposal, code design, or plan. Confidence-scored findings with constructive suggestions.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Scan for stale memory topics, near-dead instincts, and low-salience K-LEAN entries. Proposes cleanup actions interactively.
Quality gate for instinct persistence. Validates new instincts through a 4-stage pipeline before allowing them to be written to disk. Invoke before writing any new instinct file.
Use when the user needs thorough research on a complex topic, says 'deep dive', 'research this thoroughly', or 'sensor sweep'. Parallel agent orchestrator for comprehensive multi-angle investigation.
Autonomous experiment loop: iteratively improves a single file against a measurable metric. Uses git-as-state-management in a disposable worktree. Invoke when user wants to optimize, tune, or systematically improve a file through repeated iterations.
Use when you need multi-faction review combining security, strategy, cost, and quality perspectives. Trigger: fleet review, multi-faction assessment, comprehensive quality gate, or when escalation requires L5+.
Two-loop research orchestrator: chains deep-research (outer, hypothesis generation) with experiment-loop (inner, empirical testing). Use when research questions have a testable dimension with a measurable metric.
| name | opponents-view |
| description | Invoke the Devil's Advocate for steel-man + 10-dimension attack on any proposal, code design, or plan. Confidence-scored findings with constructive suggestions. |
| tags | ["quality","review","adversarial"] |
Invoke the Devil's Advocate agent to rigorously challenge a proposal, design, or code decision.
/opponents-view "<proposal>" — Challenge a proposal described inline/opponents-view --file <path> — Challenge a plan or design document/opponents-view --code <path> — Challenge code design decisions (not line-by-line review)/opponents-view --brief — Abbreviated output (top 5 findings only)If --file or --code is specified, read the target file(s).
Also check for relevant context:
mcp__memory__search_nodes) for prior decisions on this topicSpawn a Task with subagent_type: "general-purpose" using the devils-advocate agent profile.
Construct the prompt:
You are acting as the Devil's Advocate. Follow the protocol in your agent definition.
PROPOSAL:
{proposal text or file contents}
CONTEXT (prior decisions):
{knowledge graph results, if any}
Challenge this thoroughly. Steel-man first, then attack from all 10 dimensions.
Format the agent's output as:
## Opponent's View — {timestamp}
**Target**: {proposal summary}
**Verdict**: {DEFENSIBLE / MIXED / VULNERABLE}
### Steel-Man
{strongest version}
### Findings
| ID | Finding | S | Confidence | Dimension | Action |
|----|---------|---|------------|-----------|--------|
| O-001 | {description} | S0 | {0.0-1.0} | {dimension} | {suggestion} |
| O-002 | {description} | S2 | {0.0-1.0} | {dimension} | {suggestion} |
Severity: S0=Critical (conf>=0.8), S1=High (conf>=0.7), S2=Medium (conf 0.5-0.7), S3=Low (conf<0.5).
### Must-Fix Checklist
> Items at S0/S1 that need immediate attention.
- [ ] O-001: {one-line summary}
### Summary
- {N} S0 / {N} S1 / {N} S2 / {N} S3 findings
- Top recommendation: {most impactful suggestion}
### Verdict Criteria
- **DEFENSIBLE**: 0 S0, 0 S1 — proposal withstands scrutiny
- **MIXED**: 0 S0, 1-2 S1 — proposal has addressable weaknesses
- **VULNERABLE**: 1+ S0 OR 3+ S1 — proposal needs significant revision
If --brief flag is set, show only the top 5 findings (highest confidence first) and skip the steel-man section.
/codex review instead