find-next-crypto-idea
Skill178 starsChanged 4 months agoMITView on GitHub
Interview users sharply to discover, rank, or validate what they should build in crypto. Use when a user asks what to build in crypto, wants startup ideas in a crypto niche such as DeFi or AI x crypto, wants blunt feedback on an existing crypto idea, or wants a concrete artifact comparing the best next ideas. Treat the bundled idea datasets as inspiration, not constraints, and always combine them with fresh market research.
- Sends data out
In this repository34 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- animation-reverse-engineeringSkill
- brand-designSkill
- build-data-pipelineSkill
- build-defi-protocolSkill
- build-mobileSkill
- build-with-claudeSkill
- csoSkill
- debug-programSkill
- design-tasteSkill
- frontend-design-guidelinesSkill
- launch-tokenSkill
- navigate-skillsSkill
- number-formattingSkill
- page-load-animationsSkill
- product-reviewSkill
- review-and-iterateSkill
- roast-my-productSkill
- scaffold-projectSkill
- verify-humanity-pohSkill
- virtual-solana-incubatorSkill
- colosseum-copilotSkill
- competitive-landscapeSkill
- defillama-researchSkill
- learnSkill
- solana-beginnerSkill
- validate-ideaSkill
- apply-grantSkill
- create-pitch-deckSkill
- deploy-to-mainnetSkill
- marketing-videoSkill
- submit-to-hackathonSkill
- video-craftSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
26 rules. About 0 are common, 18 are distinctive, and 8 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈18
Distinctive · estimate
This author's own choices. Worth a look.
8
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 26
If TEL_PROMPTED is no: Before starting the skill workflow, ask the user about telemetry.
- Distinctive: line 36
If A: run this bash:
- Distinctive: line 43
If B: run this bash:
- Very specific: line 63
Ask the user to pick a report theme: Read references/report-themes.md and present the 12-theme table.
- Distinctive: line 64
Read references/source-map.md first to understand the local datasets available to this skill.
- Distinctive: line 65
Read references/interview-framework.md before asking substantive questions.
- Distinctive: line 68
Run fresh research with references/research-playbook.md.
- Distinctive: line 74
Keep interviewing until you can clearly state:
- Distinctive: line 80
Do not praise a bad idea because the user is attached to it.
- Distinctive: line 81
Do not produce three near-identical ideas.
- Distinctive: line 82
Always do fresh research for competitors, substitutes, and active OSS before committing to the final ranking.
- Distinctive: line 83
Always write a local HTML file.
- Distinctive: line 87
Start with anchor questions, not a giant questionnaire.
- Distinctive: line 89
Prefer questions that reveal:
- Distinctive: line 95
If the user already has an idea, stress-test it immediately.
- Distinctive: line 96
If, after several exchanges, there is still no edge, no constraint, and no credible crypto touchpoint, say so plainly and narrow the search.
- Distinctive: line 106
Treat the local datasets as archetype libraries.
- Distinctive: line 111
Read references/output-spec.md before writing the report payload.
- Very specific: line 112
If you need a structured research pass before ranking, read references/research-output-schema.md and run scripts/live_research.py.
- Very specific: line 113
Use scripts/render_report.py to render the HTML from a JSON payload.
- Distinctive: line 114
Write reports into the current working directory unless the user asks for a different path.
- Distinctive: line 115
Use file names like:
- Very specific: line 123
Write .superstack/idea-context.md in the project workspace with:
- Very specific: line 145
Fresh ideas vs. validate existing?: If user has an idea already, redirect to validate-idea skill instead.
- Very specific: line 148
Hackathon data?: If the user explicitly asks for Colosseum hackathon project data, winner patterns, or gap analysis, suggest …
- Very specific: line 175
After the skill workflow completes (success, error, or abort), log the telemetry event.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Preamble (run first)
_TEL_TIER=$(cat ~/.superstack/config.json 2>/dev/null | grep -o '"telemetryTier": *"[^"]*"' | head -1 | sed 's/.*"telemetryTier": *"//;s/"$//' || echo "anonymous")
_TEL_TIER="${_TEL_TIER:-anonymous}"
_TEL_PROMPTED=$([ -f ~/.superstack/.telemetry-prompted ] && echo "yes" || echo "no")
_TEL_START=$(date +%s)
_SESSION_ID="$$-$(date +%s)"
mkdir -p ~/.superstack
echo "TELEMETRY: $_TEL_TIER"
echo "TEL_PROMPTED: $_TEL_PROMPTED"
if [ "$_TEL_TIER" != "off" ]; then
_TEL_EVENT='{"skill":"find-next-crypto-idea","phase":"idea","event":"started","ts":"'$(date -u +%Y-%m-%dT%H:%M:%SZ)'"}'
echo "$_TEL_EVENT" >> ~/.superstack/telemetry.jsonl 2>/dev/null || true
_CONVEX_URL=$(cat ~/.superstack/config.json 2>/dev/null | grep -o '"convexUrl":"[^"]*"' | head -1 | cut -d'"' -f4 || echo "")
[ -n "$_CONVEX_URL" ] && curl -s -X POST "$_CONVEX_URL/api/mutation" -H "Content-Type: application/json" -d '{"path":"telemetry:track","args":{"skill":"find-next-crypto-idea","phase":"idea","status":"success","version":"0.2.0","platform":"'$(uname -s)-$(uname -m)'","timestamp":'$(date +%s)000'}}' >/dev/null 2>&1 &
true
fiIf TEL_PROMPTED is no: Before starting the skill workflow, ask the user about telemetry.
Use AskUserQuestion:
Help superstack get better! We track which skills get used and how long they take — no code, no file paths, no PII. Change anytime in
~/.superstack/config.json.
Options:
- A) Sure, help superstack improve (anonymous)
- B) No thanks
If A: run this bash:
echo '{"telemetryTier":"anonymous"}' > ~/.superstack/config.json
_TEL_TIER="anonymous"
touch ~/.superstack/.telemetry-promptedIf B: run this bash:
echo '{"telemetryTier":"off"}' > ~/.superstack/config.json
_TEL_TIER="off"
touch ~/.superstack/.telemetry-promptedThis only happens once. If TEL_PROMPTED is yes, skip this entirely and proceed to the skill workflow.
Wrong skill? See SKILL_ROUTER.md for all available skills.
Find Next Crypto Idea
Overview
Interview the user until there is real clarity, not just enthusiasm. Generate three serious idea candidates, force a winner, explain why crypto is necessary, and write a local HTML artifact that the user can review outside the chat.
Workflow
- Infer whether the user needs fresh ideas, a narrowed idea search in a known domain, or validation of an existing thesis.
- Ask the user to pick a report theme. Read references/report-themes.md and present the 12-theme table. Store their choice as the
themefield in the JSON payload. If they say "surprise me", pick one that matches their vibe. - Read references/source-map.md first to understand the local datasets available to this skill.
- Read references/interview-framework.md before asking substantive questions.
- Gate every candidate through references/crypto-necessity-test.md before scoring it.
- Score surviving candidates with references/scoring-rubric.md.
- Run fresh research with references/research-playbook.md. When you already have candidate ideas, use
scripts/live_research.pyto generate browser-first X queries, GitHub queries, a normalized research pack, and a worksheet HTML. - Produce the shortlist artifact first. Let the user pick one. Then deepen the chosen idea and write a second artifact.
Non-Negotiables
- Stay blunt. Challenge weak assumptions before elaborating them.
- Keep interviewing until you can clearly state:
- the user's unfair edge
- the real shipping constraint
- the first plausible wedge
- why crypto is required
- Reject ornamental crypto. Redirect to a stronger crypto angle instead of dressing up a weak one.
- Do not praise a bad idea because the user is attached to it.
- Do not produce three near-identical ideas. Force diversity across the shortlist when possible.
- Always do fresh research for competitors, substitutes, and active OSS before committing to the final ranking.
- Always write a local HTML file. Do not leave the result only in chat.
Interview Rules
- Start with anchor questions, not a giant questionnaire.
- Pull constraints only when they would change the recommendation.
- Prefer questions that reveal:
- edge
- urgency
- customer access
- shipping ability
- tolerance for infra, regulation, and sales friction
- If the user already has an idea, stress-test it immediately. Ask what breaks if the blockchain is removed.
- If, after several exchanges, there is still no edge, no constraint, and no credible crypto touchpoint, say so plainly and narrow the search.
Shortlist Rules
- Produce exactly three serious candidates unless the user explicitly asks for more.
- Include:
- a recommended winner
- why the winner wins
- why the other two lost
- a bear case for each idea
- Treat the local datasets as archetype libraries. Combine, mutate, or discard them freely if the user's context demands it.
- Integration-first framing: For each idea, note whether it can be built by integrating existing protocols or requires custom on-chain logic. Ideas that can integrate ship faster and inherit audited security — note this as context, but don't bias scoring against ideas that legitimately need a new protocol. See
data/solana-knowledge/04-protocols-and-sdks.md→ "Integrate First, Build Second" and the Decision Quick Reference for current protocol mapping.
Artifact Rules
- Read references/output-spec.md before writing the report payload.
- If you need a structured research pass before ranking, read references/research-output-schema.md and run
scripts/live_research.py. - Use
scripts/render_report.pyto render the HTML from a JSON payload. - Write reports into the current working directory unless the user asks for a different path.
- Use file names like:
idea-shortlist-YYYYMMDD-HHMMSS.htmlidea-deep-dive-YYYYMMDD-HHMMSS.html
Phase Handoff
This skill is Phase 1 (Idea) in the Idea → Build → Launch journey. After the user picks a winner from the shortlist:
- Write
.superstack/idea-context.mdin the project workspace with:phase:"idea"completed_at: ISO timestampchosen_idea: slug, name, one_liner, why_crypto, scores, competitors, mvp_checklist, gtmsource_reports: list of generated HTML filenames
- Tell the user they can proceed to the Build phase next:
scaffold-project— set up workspace with the right stackbuild-with-claude— guided MVP implementation
- See
../../../data/specs/phase-handoff.mdfor the full JSON contract.
Quick Start
# Just ask naturally — this skill activates automatically
# Example prompts:
# "What should I build in crypto?"
# "I'm interested in DeFi x AI — what are the gaps?"
# "Give me 3 Solana project ideas for a hackathon"Decision Points
- Fresh ideas vs. validate existing? If user has an idea already, redirect to
validate-ideaskill instead. - Which niche? See
../../data/ideas/for 114+ curated ideas from YC, a16z, Alliance, Superteam. - DeFi-specific? Redirect to
defillama-researchskill for data-driven DeFi discovery. - Hackathon data? If the user explicitly asks for Colosseum hackathon project data, winner patterns, or gap analysis, suggest
/colosseum-copilot. It requires a free PAT from https://arena.colosseum.org/copilot — only mention if the user asks or opts in.
Resources
references/
- references/source-map.md
- references/interview-framework.md
- references/crypto-necessity-test.md
- references/scoring-rubric.md
- references/research-playbook.md
- references/research-output-schema.md
- references/output-spec.md
- references/report-themes.md
scripts/
scripts/render_report.pyrenders the final HTML artifact from structured JSON.scripts/live_research.pygenerates browser-first research plans, query URLs, a normalizedresearch-pack.json, and an optional worksheet HTML.
assets/
assets/report-template.htmlis the HTML template used by the render script.assets/research-worksheet-template.htmlis the HTML template for the clickable research worksheet.
Telemetry (run last)
After the skill workflow completes (success, error, or abort), log the telemetry event.
Determine the outcome from the workflow result: success if completed normally, error
if it failed, abort if the user interrupted.
Run this bash:
_TEL_END=$(date +%s)
_TEL_DUR=$(( _TEL_END - ${_TEL_START:-$_TEL_END} ))
_TEL_TIER=$(cat ~/.superstack/config.json 2>/dev/null | grep -o '"telemetryTier": *"[^"]*"' | head -1 | sed 's/.*"telemetryTier": *"//;s/"$//' || echo "anonymous")
if [ "$_TEL_TIER" != "off" ]; then
echo '{"skill":"find-next-crypto-idea","phase":"idea","event":"completed","outcome":"OUTCOME","duration_s":"'"$_TEL_DUR"'","session":"'"$_SESSION_ID"'","ts":"'$(date -u +%Y-%m-%dT%H:%M:%SZ)'","platform":"'$(uname -s)-$(uname -m)'"}' >> ~/.superstack/telemetry.jsonl 2>/dev/null || true
true
fiReplace OUTCOME with success/error/abort based on the workflow result.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

