How do I install the Consult Red Team skill?

Run npx skillmds@latest add timelnapp/consult-red-team in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Consult Red Team safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Consult Red Team?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Consult Red Team free to use?

Yes. Installing skills from SkillMD is free. This skill is licensed under MIT.

Who published Consult Red Team?

Timelnapp (@timelnapp) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Consult Red Team by timelnapp — Agent Skill

Name: Consult Red Team
Author: Timelnapp

Consult Red Team

Trigger on "red team proposal", "devil's advocate proposal", "stress test proposal", "consult red team", "win themes challenge", "cos red team". Multi-perspective review: client exec, procurement, technical skeptic, devil's advocate. Feeds timeln-warned scars into DA. NOT lint (consult-consistency-lint), NOT integrity (consult-integrity).

Timelnapp cab23ce 2.5 KB Updated Sep 17, 2026

File contents

Consult Red Team -- Multi-Perspective Proposal Review

Stress-test the engagement pack before ship. Anti-sycophancy: Devil's Advocate concedes only when severity ≥ 4 (1–5 scale).

Inputs

Input	Required
Latest frame through commercial + architecture pointer	yes
`lint-report.md` if exists	recommended
Framework variant index + chosen merge	recommended
`timeln-warned` on primary technical bet	run before review

Prefetch (mandatory)

timeln-warned — topic = primary technical path from arc (e.g. "GPU fine-tuning FSI transcripts")
Pass top scars into Devil's Advocate section verbatim

Four perspectives

Perspective	Question	Output
Client exec	Can I say yes/no in one read?	Score 0–100 + 3 bullets
Procurement	Are assumptions, out-of-scope, and commercial gaps visible?	Findings table
Technical skeptic	Does architecture + acceptance hold under audit?	Findings table
Devil's advocate	What kills this deal? Use timeln-warned scars. Concede only if severity ≥ 4	BLOCK / ALLOW + severity

Scoring rubric for exec perspective: reuse ../consult-arc/FRAMEWORK-VARIANTS.md criteria (decision clarity, measurability, commercial defensibility, technical credibility) — average × 20 for 0–100.

Verdict

BLOCK — DA severity ≥ 4 on any show-stopper, OR exec score < 60
ALLOW — otherwise; warnings ship with pack

Output -- exactly this shape

## Red team -- <client> -- <YYYY-MM-DD>

**Verdict:** BLOCK | ALLOW
**Exec score:** <0-100>

### Client exec
- ...
**Score rationale:** ...

### Procurement
| Finding | Severity | Suggested fix |
|---|---|---|

### Technical skeptic
| Finding | Severity | Suggested fix |
|---|---|---|

### Devil's advocate
**Severity (1-5):** n
**Scars from timeln-warned:**
- ...
**Challenge:** ...
**Concede?** yes (≥4) | no

### Merge recommendation
<one line if variants exist>

Rules

Do not resolve lint/integrity conflicts — reference finding IDs.
DA must cite timeln-warned when available; if no record, say so — do not invent failures.
BLOCK triggers pipeline revision loop (max 2).

---name: consult-red-teamdescription: Trigger on "red team proposal", "devil's advocate proposal", "stress test proposal", "consult red team", "win themes challenge", "cos red team". Multi-perspective review: client exec, procurement, technical skeptic, devil's advocate. Feeds timeln-warned scars into DA. NOT lint (consult-consistency-lint), NOT integrity (consult-integrity).license: MIT---# Consult Red Team -- Multi-Perspective Proposal ReviewStress-test the engagement pack before ship. Anti-sycophancy: Devil's Advocate concedes only when severity ≥ 4 (1–5 scale).## Inputs| Input | Required ||---|---|| Latest frame through commercial + architecture pointer | yes || `lint-report.md` if exists | recommended || Framework variant index + chosen merge | recommended || **`timeln-warned`** on primary technical bet | run before review |## Prefetch (mandatory)1. **`timeln-warned`** — topic = primary technical path from arc (e.g. "GPU fine-tuning FSI transcripts")2. Pass top scars into Devil's Advocate section verbatim## Four perspectives| Perspective | Question | Output ||---|---|---|| **Client exec** | Can I say yes/no in one read? | Score 0–100 + 3 bullets || **Procurement** | Are assumptions, out-of-scope, and commercial gaps visible? | Findings table || **Technical skeptic** | Does architecture + acceptance hold under audit? | Findings table || **Devil's advocate** | What kills this deal? Use timeln-warned scars. Concede only if severity ≥ 4 | BLOCK / ALLOW + severity |Scoring rubric for exec perspective: reuse `../consult-arc/FRAMEWORK-VARIANTS.md` criteria (decision clarity, measurability, commercial defensibility, technical credibility) — average × 20 for 0–100.## Verdict- **BLOCK** — DA severity ≥ 4 on any show-stopper, OR exec score < 60- **ALLOW** — otherwise; warnings ship with pack## Output -- exactly this shape```## Red team -- <client> -- <YYYY-MM-DD>**Verdict:** BLOCK | ALLOW**Exec score:** <0-100>### Client exec- ...**Score rationale:** ...### Procurement| Finding | Severity | Suggested fix ||---|---|---|### Technical skeptic| Finding | Severity | Suggested fix ||---|---|---|### Devil's advocate**Severity (1-5):** n**Scars from timeln-warned:**- ...**Challenge:** ...**Concede?** yes (≥4) | no### Merge recommendation<one line if variants exist>```## Rules- Do not resolve lint/integrity conflicts — reference finding IDs.- DA must cite timeln-warned when available; if no record, say so — do not invent failures.- BLOCK triggers pipeline revision loop (max 2).

Frequently asked questions