grill-me-well-done-generic - SKILL.md Agent Skill

name: grill-me-well-done-generic description: > Use this skill when the user provides any plan, idea, document, draft, strategy, research proposal, content plan, business plan, technical plan, creative concept, learning plan, or project outline that needs to be challenged, improved, and made more decision-ready. The skill dynamically identifies the domain, required expertise, reviewer personas, evaluation criteria, risks, alternatives, and produces an enhanced version of the original document. metadata: version: "1.0.0"

Grill-Me-Well-Done Generic Skill

Purpose

This skill transforms an initial plan, idea, or document into a stronger, clearer, more complete, more realistic, and more defensible version.

It works across domains, including but not limited to:

engineering
research
business
education
content creation
product strategy
marketing
creative writing
operations
personal planning
academic work
entrepreneurship
public communication
policy
design
finance
career planning

The skill must not assume the domain in advance.

It must infer the required expertise from the user’s input and, when needed, ask a short interview to identify the correct reviewer personas and evaluation criteria.

Core Principle

Do not simply rewrite the document.

First understand it, challenge it, improve it, and only then produce the enhanced version.

The skill behaves like a custom expert review board assembled specifically for the provided document.

Inputs

The user may provide:

a rough idea
a plan
a document
a draft
an outline
a proposal
a strategy
a research question
a content plan
a business concept
a technical design
a creative concept
notes or bullet points

If the input is incomplete, proceed with best effort and mark assumptions clearly.

Phase 1 — Intake and Domain Discovery

Analyze the provided material and identify:

document type
domain or field
intended audience
purpose
desired outcome
level of maturity
decision context
risks if the document is wrong
missing information
implicit assumptions
required expertise areas

If the domain, audience, or success criteria are unclear, ask a short interview.

Phase 2 — Short Interview

Ask only the minimum questions needed to assemble the right review panel.

Default interview questions:

Who is the final audience for this document?
What does success look like?
What are the main constraints?
What level of challenge do you want: light, serious, or brutal?
Should the final output be strategic, practical, academic, creative, executive, or implementation-ready?

Optional questions depending on context:

What is the deadline?
What resources are available?
What risks worry you most?
Are there any non-negotiables?
Should the output preserve the original tone or fully restructure it?

If the user does not answer, infer reasonable assumptions and continue.

Phase 3 — Dynamic Expert Panel Creation

Create reviewer personas based on the document.

Do not use a fixed set of personas.

Instead, infer the right ones from the content.

Examples:

For a research proposal:

Methodology Reviewer
Literature Review Reviewer
Statistical Validity Reviewer
Ethics Reviewer
Publication Reviewer

For a content creator plan:

Audience Strategist
Narrative Reviewer
Platform Growth Reviewer
Brand Consistency Reviewer
Monetization Reviewer

For a business plan:

Market Analyst
Financial Reviewer
Operations Reviewer
Customer Acquisition Reviewer
Risk Reviewer

For an engineering plan:

Domain Expert
Safety Reviewer
Cost Reviewer
Reliability Reviewer
Implementation Reviewer

For a personal career plan:

Career Strategist
Skills Gap Reviewer
Market Reality Reviewer
Execution Coach
Risk Reviewer

Always include at least:

Domain Expert
Skeptical Reviewer
Practical Execution Reviewer
Risk Reviewer
Audience/User/Stakeholder Reviewer

Add specialized reviewers only when the document requires them.

Phase 4 — Dynamic Evaluation Criteria

Create evaluation criteria based on the document’s purpose.

Possible criteria include:

clarity
completeness
feasibility
originality
evidence quality
audience fit
strategic value
practical usefulness
risk awareness
ethical soundness
cost realism
time realism
operational readiness
creativity
persuasiveness
accuracy
implementation readiness
scalability
maintainability
scientific validity
emotional impact
market fit
differentiation
consistency
decision quality

Select only the criteria relevant to the provided document.

Do not over-evaluate simple documents.

Phase 5 — Tree-of-Thoughts Review

Use branching review internally.

Explore several possible improvement paths:

Root: User-provided document
  ├── Path A: Improve clarity
  ├── Path B: Improve correctness
  ├── Path C: Improve feasibility
  ├── Path D: Improve impact
  ├── Path E: Reduce risk
  ├── Path F: Improve audience fit
  ├── Path G: Challenge the core direction
  └── Path H: Explore stronger alternatives

For each relevant path:

Identify weaknesses.
Identify hidden assumptions.
Ask what could fail.
Consider alternatives.
Decide whether to improve, replace, or preserve the original idea.

Do not expose private reasoning chains.

Expose only useful conclusions.

Phase 6 — Grill-Me Challenge

Challenge the document through the dynamically created personas.

Each persona should answer:

What is strong?
What is weak?
What is missing?
What could fail?
What assumption is dangerous?
What would make this more credible?
What would make this more useful?
What should be removed?
What should be added?
What alternative should be considered?

The challenge must be direct but constructive.

The goal is improvement, not criticism for its own sake.

Phase 7 — Scoring

Score the current document only on relevant criteria.

Use a 1–5 scale.

Example:

Clarity: 3/5
Feasibility: 2/5
Audience Fit: 4/5
Risk Awareness: 2/5
Evidence Quality: 3/5
Execution Readiness: 2/5

Then summarize:

strongest elements
weakest elements
top risks
missing decisions
highest-impact improvements

Phase 8 — Alternative Options

Generate alternatives only when useful.

Possible alternatives:

simpler version
more ambitious version
lower-risk version
lower-cost version
more creative version
more evidence-based version
more audience-focused version
more execution-ready version

Do not force alternatives when the document is already narrow and clear.

Phase 9 — Enhanced Document

Rewrite or enhance the original document.

The enhanced version may include:

title
objective
context
target audience
problem statement
desired outcome
scope
non-goals
assumptions
constraints
proposed approach
alternatives considered
risks
mitigations
evidence needed
execution plan
success criteria
open questions
final recommendation

Use only the sections that make sense for the document type.

Do not make every output look like a software architecture document.

Adapt the structure to the domain.

Phase 10 — Final Output Format

Default response structure:

1. Intake Summary
2. Expert Panel Created
3. Evaluation Criteria
4. Grill-Me Findings
5. Risk & Gap Analysis
6. Alternative Options
7. Enhanced Document
8. Final Recommendation
9. Open Questions

If the user asks for a clean final document only, provide only:

Enhanced Document

Behavior Rules

Be field-agnostic.
Infer expertise dynamically.
Ask only necessary questions.
Do not overcomplicate simple plans.
Do not invent facts.
Mark assumptions clearly.
Preserve useful parts of the original document.
Remove weak, redundant, or unsupported parts.
Challenge premature conclusions.
Improve both content and structure.
Adapt tone to the intended audience.
Produce something the user can actually use.

Quality Bar

The final enhanced document should be suitable for one of the following:

stakeholder review
expert review
publishing
execution
decision-making
research refinement
creative development
business validation
personal planning
academic review

The final result must be stronger than the original in clarity, usefulness, credibility, and decision quality.