Skip to content
Back to skills

Perspective Council

ASecurity

Use only when the user explicitly requests a perspective council or war room. Answer ordinary decisions and critique requests without a council.

  • 2 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added October 15, 2026
research

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned October 15, 2026

npx -y skills add MVPavan/via --skill perspective-council --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Perspective Council?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Perspective Council
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/mvpavan-perspective-council/badge)](https://www.skillsdirectory.com/skills/mvpavan-perspective-council)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: perspective-council
description: Use only when the user explicitly requests a perspective council or war room. Answer ordinary decisions and critique requests without a council.
---

# Perspective Council

An explicitly requested council examines a consequential decision through distinct
lenses. Named model members and a judge belong to model-council instead.

## Choose the declared scale

For a focused council, select a small set of relevant opposing lenses and one
fresh synthesis pass. State membership and whether peer review is included before
dispatch. Full mode is five advisors, five anonymous peer reviews, then one chair;
use it when explicitly requested or covered by the agreed scope and budget.

| Lens | Contribution |
|---|---|
| Contrarian | Failure modes and neglected downside |
| First Principles | Actual problem and binding assumptions |
| Expansionist | Plausible upside and overlooked opportunity |
| Outsider | Missing explanations and assumed context |
| Executor | Feasibility and a concrete first step |

Frame the question neutrally with relevant facts, constraints and uncertainty.
Each advisor gets the same grounding plus its lens, and no other advisor output.
Sequential dispatch can preserve independence; simultaneity is not a requirement.
Require evidence and calibrated uncertainty rather than forcing certainty or
ignoring risks to perform a role.

When peer review is included, anonymize responses before reviewers compare their
reasoning and gaps. Give the fresh chair all responses and peer reviews, restoring
attribution where helpful. Deliver agreement, disagreements, blind spots,
recommendation and a practical next step. Agreement among agents is not
independent factual verification or automatic high confidence.

Use permitted configured agents and shared delegation policy. Report unavailable
passes honestly. Save a transcript only when requested; a council does not grant
implementation or external-message authority.

Files in this skill

  • SKILL.md1.9 KB
  • agents/openai.yaml42 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.