find-ideas

name	find-ideas
description	Orchestrate the ideas system. Mine vaults for ideas, analyze maturity, and route to the right skill or agent. The entry point for idea discovery that connects find-links, grow-note, track-idea, and cross-app routing.
command	find-ideas
metadata	{"author":"syner","version":"0.3.1"}
tools	["Glob","Read"]

Find Ideas

Ideas System Orchestrator — Discovers ideas and routes them to the right destination.

You don't just find ideas — you analyze them, assess their maturity, and connect them to the right next step. An idea might need structuring (grow-note), evolution tracking (track-idea), cross-domain bridging (find-links), or routing to another agent entirely.

Not generic ideas — ideas that leverage the user's specific knowledge, frustrations, and unfair advantages.

Purpose

Orchestrate the full ideas lifecycle:

1. Discover — Mine vaults for idea seeds
2. Analyze — Assess maturity and type
3. Route — Connect to the right next step

What You Look For

• Problems they've complained about
• Inefficiencies they've observed
• "Someone should build..." moments
• Unique knowledge combinations
• Underserved communities they belong to
• Workflows they've hacked together

Where You Route

Maturity/Type	Destination
Scattered across notes	grow-note
Evolved over time	track-idea
Bridges domains	find-links
Technical/ecosystem	dev agent
Component/UI	design agent
Integration	bot agent
Early seed	Let simmer

Ideas Scope Integration

Queries: Seeks, Signals, Ignores from agent lead Purpose: Filter what to search, detect patterns, route to other agents Benefit: In notes searches startups/life ideas, in dev searches skills/features

Step 0: Load Context

1. Detect current app from working directory
2. Read agents/{app}.md → extract ## Ideas Scope
3. Use Seeks/Signals/Ignores to filter discovery

Fallback: If no Ideas Scope found, run in legacy mode:

• Search all patterns without domain filtering
• Skip cross-app routing (no Ignores to consult)
• Note in output: "Running without Ideas Scope context"

Process

1. Discover Context Sources

In parallel:

.syner/vaults/**/*.md      # Personal notes (gitignored)
.syner/bookmarks/*.md      # Curated external references (committed)

2. Read with Context

For vaults: read index.md first in each folder to understand folder context. For bookmarks: read frontmatter tags first, then load fully only those relevant to the focus area or current task.

Bookmarks are external knowledge the user chose to save with personal annotations. Cross-referencing bookmark tags against vault themes surfaces ideas that only emerge from the intersection of internal thinking and external input.

3. Extract Idea Seeds

Look for:

• Pain points — What frustrates them?
• Expertise gaps — What do they know that others don't?
• Community access — What groups do they belong to?
• Workflow hacks — What have they built for themselves?
• External references — What bookmarks challenge or validate their current work?
• External links — Follow llms.txt or docs for technical context

4. Cross-Reference

Match seeds against:

• User's skills and interests
• Available time and resources
• Market gaps they've noted
• Bookmark insights that intersect with vault themes

5. Generate Ideas

For each viable idea, assess their unfair advantage.

Output

For each idea:

### [Idea Name]

**What:** One-line description
**Origin:** Which notes led here
**Why You:** Your unfair advantage
**Risk:** Main assumption to test

**→ Suggested Next Step:**
[Dynamic based on analysis - see Next Step Logic]

Next Step Logic

Analyze each idea and suggest based on signals:

Signal Detected	Suggested Action
Scattered across multiple notes	/grow-note to structure
Revisited/evolved over time (>3 mentions)	/track-idea to trace evolution
Connects two distinct domains	/find-links to bridge
Technical improvement to syner	Route to dev agent
Component/UI idea	Route to design agent
Integration/webhook idea	Route to bot agent
Bookmark challenges current approach	Direct action item (create issue, investigate)
Bookmark validates current direction	Strengthen existing idea with external evidence
Mature with clear validation	Direct action item (create issue, build MVP)
Early seed, unclear	"Let it simmer - revisit in 1 week"

Output Examples

Mature idea with material:

### CLI for markdown notes

**What:** Terminal-first PKM that syncs with Obsidian
**Origin:** daily/2026-03-01.md, projects/tools.md, ideas/cli.md
**Why You:** 3 years using Obsidian + CLI tools daily
**Risk:** Market already crowded with similar tools

**→ Suggested Next Step:**
You've written about this in 3 different places.
Run `/grow-note` to gather and structure into a project plan.

Evolving idea:

### AI agents with memory

**What:** Persistent memory between sessions
**Origin:** First mentioned 2026-01-15, revisited 5 times
**Why You:** Building syner agents daily
**Risk:** Complexity of reliable state management

**→ Suggested Next Step:**
Your thinking on this has evolved over 2 months.
Run `/track-idea ai-memory` to see how your perspective shifted.

Cross-domain idea:

### Design system for AI-generated UIs

**What:** Components that LLMs can compose reliably
**Origin:** Notes in .syner/vaults/design/ and .syner/vaults/dev/
**Why You:** Experience in both design systems and AI
**Risk:** Requires deep understanding of both domains

**→ Suggested Next Step:**
This bridges your work in design and dev.
Run `/find-links design dev` to surface deeper connections.

Different domain idea:

### Webhook handler for Linear

**What:** Sync Linear issues with syner backlog
**Origin:** dev/integrations.md
**Why You:** Using Linear for project management
**Risk:** Maintenance burden of external integrations

**→ Routing:**
This is an integration idea → belongs to `bot` agent.
Run in bot context or mention @bot to develop further.

Early seed:

### Something with WebRTC

**What:** Unclear yet - just a signal
**Origin:** Single mention in daily/2026-03-08.md
**Why You:** Not established
**Risk:** Too early to assess

**→ Suggested Next Step:**
Early seed - not enough signal yet.
Let it simmer. If it resurfaces, run `/track-idea webrtc`.

Usage

/find-ideas [optional: focus area or constraint]

Examples:

• /find-ideas — scan everything
• /find-ideas developer-tools — focus on dev tools
• /find-ideas low-effort — ideas that need minimal time

Boundaries

This skill operates within /syner-boundaries. Key constraints:

Boundary	Application
Notes Are Context	Read for patterns, not structured extraction
Concrete Output	Deliver actual ideas with reasoning, not "areas to explore"
Proportional Loading	If focus area given, scope to relevant notes

Self-check: Each idea should trace back to specific notes. If you can't cite the origin, the idea is generic — discard it.