From ideation-copilot
Score a business idea by dispatching evaluation agents that research and rate the idea across investment and market dimensions. Produces a structured, machine-readable report.
How this skill is triggered — by the user, by Claude, or both
Slash command
/ideation-copilot:idea-evaluateThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Score a business idea by dispatching evaluation agents that research and rate the idea across investment and market dimensions. Produces a structured, machine-readable report.
Score a business idea by dispatching evaluation agents that research and rate the idea across investment and market dimensions. Produces a structured, machine-readable report.
Follow these phases in order.
<argument> for the idea folder name and an optional agent filter (vc or market)ideas/*{idea-folder-name}*/
references/evaluation-framework.md
references/exa-research.md
03-assumptions.md exists, flag it for cross-referencing.Prepare the context block for agents — combine:
If an agent filter was specified (vc or market), run only that agent.
Otherwise, dispatch both agents in parallel:
Launch two Agent tasks simultaneously, each receiving the full context block:
Each agent:
Collect outputs from both agents (or one if filtered). Compute:
If all agents ran:
combined_score = round((vc_overall + market_overall + yc_overall) / 3)
If two agents ran: average those two scores. If single agent: use that agent's overall score as the combined score.
Identify:
Present a summary table:
## Evaluation Summary
| Agent | Score | Deal-breakers |
|-------|-------|---------------|
| VC Investability | [score]/100 | [list or None] |
| Market Opportunity | [score]/100 | — |
| YC Founder-Fit | [score]/100 | — |
| **Combined** | **[score]/100** | **[count] deal-breaker(s)** |
**Weakest dimension:** [agent] → [dimension] ([score]/5) — address this first.
Then present the full per-dimension analysis from each agent.
Write the evaluation report to the idea folder:
Filename: evaluation-YYYYMMDD-HHmmss.md
Location: Inside the idea folder (ideas/{idea-name}/)
The file starts with YAML frontmatter containing all scores in a machine-readable format:
---
type: evaluation
date: YYYY-MM-DD
agents: [vc, market-analyst, yc-founder-fit]
combined_score: 59
deal_breakers: ["Team scored 1/5"]
scores:
vc:
overall: 58
dimensions:
team: { score: 2, weight: 2.0 }
timing: { score: 4, weight: 1.5 }
tam: { score: 3, weight: 1.0 }
technology: { score: 3, weight: 1.0 }
competition: { score: 2, weight: 1.0 }
business_model: { score: 3, weight: 1.0 }
gtm: { score: 2, weight: 1.0 }
traction: { score: 1, weight: 1.0 }
market_analyst:
overall: 66
dimensions:
market_size: { score: 4 }
competitive_landscape: { score: 3 }
timing_tailwinds: { score: 4 }
customer_accessibility: { score: 3 }
regulatory_risk: { score: 3 }
yc_founder_fit:
overall: 54
dimensions:
founder_market_fit: { score: 4 }
market_size: { score: 3 }
problem_acuteness: { score: 2 }
competition_presence: { score: 3 }
personal_peer_demand: { score: 2 }
recent_possibility: { score: 4 }
successful_proxies: { score: 3 }
long_term_commitment: { score: 3 }
business_scalability: { score: 2 }
idea_space_fertility: { score: 3 }
weakest_dimension: { agent: "vc", dimension: "traction", score: 1 }
---
Below the frontmatter: the full evaluation summary table, per-agent analysis, and all per-dimension details.
After presenting the report:
"These are the assessments based on your idea docs and market research. Challenge any score you disagree with, or say 'done' to finish."
If the founder challenges a score:
Session commands:
After presenting the evaluation, suggest next steps based on the scores:
Evaluation complete! What's next?
→ /idea:update {idea-name} — fix low scores by adding missing info to your docs
→ /idea:pushback {idea-name} — stress-test the claims behind your weakest dimensions
→ /idea:forge {idea-name} — synthesize everything if you've done multiple rounds
Choose based on the scores:
/idea:evaluate my-idea vc or market or yc produces only that agent's scores. The combined score equals the single agent's score.If Exa tools were not available during this evaluation, append a single note at the end of the report:
Tip: Run
/idea:setupto configure Exa search for richer market and competitor data.
npx claudepluginhub kaminskypavel/ideation-copilot --plugin ideation-copilotProvides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.