115 lines
2.9 KiB
Markdown
115 lines
2.9 KiB
Markdown
# Intent
|
|
|
|
## Purpose
|
|
|
|
This skill helps an agent evaluate whether a user idea, MVP, product concept, or startup concept is worth pursuing.
|
|
|
|
The skill should behave like a structured decision engine:
|
|
|
|
- gather opposing viewpoints,
|
|
- pressure-test the idea,
|
|
- judge market potential,
|
|
- and give the user a clear recommendation.
|
|
|
|
## What It Should Do
|
|
|
|
When the user presents an idea, the skill should:
|
|
|
|
- research and reason through the idea from multiple angles,
|
|
- split the analysis into pro, con, and neutral positions,
|
|
- use subagents to avoid one-sided thinking,
|
|
- identify market signals, risks, and opportunity size,
|
|
- and produce a final score plus a decision summary.
|
|
|
|
## Core Question
|
|
|
|
The skill exists to answer:
|
|
|
|
- Is this a good idea?
|
|
- Is there a market for it?
|
|
- What are the strongest reasons to build it?
|
|
- What are the strongest reasons not to build it?
|
|
- Is it worth planning an MVP?
|
|
|
|
## Operating Model
|
|
|
|
The evaluation must happen in 3 stages.
|
|
|
|
Before moving to the next stage, wait until every agent in the current stage has finished its research and returned its output.
|
|
|
|
### Stage 1: Broad Scan
|
|
|
|
Launch 5 subagents.
|
|
|
|
Assign them different roles so the idea is explored from multiple directions.
|
|
|
|
Required split:
|
|
|
|
- 2 subagents against the idea
|
|
- 2 subagents for the idea
|
|
- 1 neutral subagent that gives 1 strong pro and 1 strong con
|
|
|
|
Stage 1 should feel exploratory and independent. The subagents should not all say the same thing.
|
|
|
|
### Stage 2: Deep Debate
|
|
|
|
Launch 2 additional subagents.
|
|
|
|
Required split:
|
|
|
|
- 1 fully pro
|
|
- 1 fully against
|
|
|
|
These subagents should make the strongest possible case for their side using:
|
|
|
|
- market logic,
|
|
- customer pain,
|
|
- competition,
|
|
- timing,
|
|
- feasibility,
|
|
- and evidence where possible.
|
|
|
|
Stage 2 should sharpen the argument on both sides, not repeat Stage 1.
|
|
|
|
### Stage 3: Final Judgment
|
|
|
|
The main agent reads all subagent output and makes the final call.
|
|
|
|
The final response must:
|
|
|
|
- give the idea a score from 1 to 10,
|
|
- summarize the strongest positive arguments,
|
|
- summarize the strongest negative arguments,
|
|
- state a clear verdict,
|
|
- and ask whether the user wants to plan an MVP next.
|
|
|
|
## Output Format
|
|
|
|
The final response should be concise and decision-oriented.
|
|
|
|
It should include:
|
|
|
|
- score out of 10,
|
|
- short verdict,
|
|
- positives,
|
|
- negatives,
|
|
- and an MVP question.
|
|
|
|
## Behavioral Rules
|
|
|
|
- Keep pro and con agents meaningfully different.
|
|
- Treat the neutral agent as a real balancing role, not filler.
|
|
- Prefer evidence and market reasoning over vague enthusiasm.
|
|
- If the market looks weak, crowded, or unproven, say so plainly.
|
|
- If the idea looks strong, explain why it survives criticism.
|
|
- Avoid giving the user a false sense of certainty.
|
|
|
|
## Success Criteria
|
|
|
|
The skill is successful if it can:
|
|
|
|
- expose weak ideas quickly,
|
|
- support strong ideas with real reasoning,
|
|
- surface useful market and execution risks,
|
|
- and help the user decide whether the idea deserves an MVP.
|