From ideation-copilot
Synthesize all accumulated knowledge about an idea into a consolidated summary — score trajectory, key findings, validated vs assumed, and pitch-ready overview. Use when the idea has been through multiple rounds and needs a clear picture of where it stands.
How this skill is triggered — by the user, by Claude, or both
Slash command
/ideation-copilot:idea-forgeThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Read everything accumulated about an idea — docs, evaluations, pushback sessions, predictions — and synthesize it into a single consolidated summary. This is the compounding step: it makes all the accumulated knowledge legible.
Read everything accumulated about an idea — docs, evaluations, pushback sessions, predictions — and synthesize it into a single consolidated summary. This is the compounding step: it makes all the accumulated knowledge legible.
Read <argument> to identify the idea folder. Search for it under ideas/.
ideas/*{argument}*/
Read ALL files in the folder:
evaluation-*.md files (sorted by date for trajectory)pushback-session-*.md filespushback-predictions-*.md filesParse YAML frontmatter from all evaluation files. Build a timeline:
## Score Trajectory
| Date | Combined | VC | Market | Weakest Dimension | Deal-breakers |
|------|----------|-----|--------|-------------------|---------------|
| 2026-03-19 | 42/100 | 34 | 50 | Team (1/5) | Team |
| 2026-03-20 | 58/100 | 52 | 64 | Traction (2/5) | None |
| 2026-03-22 | 74/100 | 70 | 78 | GTM (3/5) | None |
**Trend:** Improving (+32 points across 3 evaluations)
**Biggest improvement:** Team (1→4, after adding CTO background and advisor plan)
**Persistent weakness:** GTM — still no proven cold acquisition channel
If only one evaluation exists, show it as the baseline with no trajectory.
Read all pushback sessions and compile:
What's been validated:
What's been refuted or remains risky:
Key pivots and evolution:
Create forge-YYYYMMDD.md in the idea folder:
# Forge: <idea name>
Date: <date>
Evaluations analyzed: <count>
Pushback sessions analyzed: <count>
## Idea in One Paragraph
<Synthesized from 00-overview, incorporating all evolution>
## Score Trajectory
<table from Step 2>
## What's Validated
- <validated claim/assumption with source>
## What's Still Risky
- <unvalidated claim with what would test it>
## Key Pivots
1. <date>: <what changed and why>
## Open Questions
<Compiled from all sessions and docs>
## Pitch-Ready Summary
<3-5 bullet points a founder could use to pitch this idea right now, grounded in validated evidence>
## Verdict
**Idea strength:** [Strong / Promising / Needs work / Reconsider]
**Confidence level:** [High / Medium / Low] — based on ratio of validated vs assumed claims
**Biggest remaining risk:** <one sentence>
**Recommended next action:** <one specific thing to do>
Output the forge summary, then:
Forge complete! Here's where this idea stands.
→ /idea:pushback {idea-name} — if there are still untested claims to challenge
→ /idea:evaluate {idea-name} — if you've made changes since the last score
→ /idea:new {new-idea-name} — if the forge reveals a pivot worth exploring as a new idea
Provides 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.
npx claudepluginhub kaminskypavel/ideation-copilot --plugin ideation-copilot