llm-council

Run any question, idea, or decision through a council of 5 AI advisors who independently analyze it, peer-review each other anonymously, and synthesize a final verdict. Based on Karpathy's LLM Council methodology. MANDATORY TRIGGERS: 'council this', 'run the council', 'war room this', 'pressure-test this', 'stress-test this', 'debate this'. STRONG TRIGGERS (use when combined with a real decision or tradeoff): 'should I X or Y', 'which option', 'what would you do', 'is this the right move', 'validate this', 'get multiple perspectives', 'I can't decide', 'I'm torn between'. Do NOT trigger on simple yes/no questions, factual lookups, or casual 'should I' without a meaningful tradeoff (e.g. 'should I use markdown' is not a council question). DO trigger when the user presents a genuine decision with stakes, multiple options, and context that suggests they want it pressure-tested from multiple angles.

Quality

92%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

85%Weight 40%Scale 1-3

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A highly actionable, clearly sequenced skill with complete prompt templates and a worked example. Its main weakness is conciseness: motivational framing and repeated explanations of concepts Claude already knows could be trimmed to respect the context budget.

Suggestions

Trim the motivational/rationale prose in 'the five advisors' and 'Why these five' to the distinct thinking styles only, removing restatements of how the council works that are already covered in the workflow steps.

Consolidate the repeated descriptions of the synthesis structure (step 4 prose, the chairman template, and step 5 output format) into a single canonical structure to avoid threefold duplication.

Remove the opening 'You ask one AI a question...' preamble and the 'This is adapted from Andrej Karpathy...' paragraph unless the provenance is essential to execution, as these explain background rather than instruct.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient and actionable, but it pads with concept restatements Claude already knows ('The council fixes this', the 'Why these five' tensions prose, and repeated restatements of the synthesis logic) that could be tightened.	2 / 3
Actionability	Provides fully copy-paste-ready sub-agent prompt templates for all three rounds (advisor, reviewer, chairman) plus an exact markdown output structure and a worked end-to-end example.	3 / 3
Workflow Clarity	A clearly sequenced six-step process (frame → convene → peer review → synthesize → present → transcript) with explicit parallelization instructions; the operations are non-destructive so the missing-validation cap does not apply.	3 / 3
Progressive Disclosure	A self-contained single-file skill with no bundle files and clear, well-organized section headings that aid navigation; per the simple-skills scoring note, well-organized single-file content scores 3.	3 / 3
	Total	11 / 12 Passed

Description

100%Weight 40%Scale 1-3

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, well-constructed description that clearly states concrete capabilities, provides extensive natural trigger terms, and explicitly demarcates when to use versus not use the skill. It reads in third person and avoids vague fluff.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'independently analyze it, peer-review each other anonymously, and synthesize a final verdict' — matching the anchor for listing several specific concrete actions.	3 / 3
Completeness	Explicitly answers both what (the council methodology and its three phases) and when (MANDATORY TRIGGERS, STRONG TRIGGERS, and explicit negative triggers), satisfying the both-what-and-when anchor.	3 / 3
Trigger Term Quality	Strong coverage of natural phrasings users would actually say: 'council this', 'pressure-test this', 'should I X or Y', 'I can't decide', 'I'm torn between' — matching the good-coverage anchor.	3 / 3
Distinctiveness Conflict Risk	Clear niche (multi-advisor peer-review synthesis) with distinct triggers plus explicit 'Do NOT trigger on...' guidance that reduces overlap with other skills.	3 / 3
	Total	12 / 12 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: aiwithremy/claude-skills-llm-council
Path: SKILL.md
Commit: 55ee36e

Reviewed: 12 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.