AI Greenhouse Agent Tends Ideas to Ripeness

Idea-Tending Model Shifts Notes from Static Jars to Growing Gardens

Treat ideas like plants in a greenhouse: create conditions for organic growth instead of static capture. Ideas progress through 6 states—seed (isolated thought), signal (supporting evidence), seedling (planted raw), growing (attracting connections), ripening (near writable), wilting (needing decision), and composting (retired but retrievable). This model fixes overwhelmed notes apps by using a physical file system (garden/ with inbox/, seeds/, ready/, compost/) where harvesting means manually moving files from seeds/ to ready/.

Impact: Prevents dead ideas by enforcing patience—e.g., a shower thought left 18 days collects signals from client talks and readings, revealing angles instantly. Unlike Karpathy's LLM knowledge bases (for archiving consumption), this grows creation: agent reads garden-state.md index first for stats (seed count, themes, ripeness), clusters, convergence warnings (e.g., seeds from 3 months apart matching), and orphan watch, scaling efficiently without scanning all files.

Modular Rules and Skills Power Consistent Gardener Behavior

Divide agent into 11 rule files (grouped: 3 identity, 3 mechanics, 3 edges, 2 session) and 5 skills for precision over improvisation. Rules ensure patient, non-pushy voice: e.g., 04-personality.md sets observant tone; 03-scope.md blocks writing content/research/deletions; 09-user-patterns.md tracks planting frequency, theme dominance, germination response rate, ripeness action speed to adapt (shorten questions if ignored, prioritize cross-refs in bursts).

5 skills trigger via 4 commands:

🪴 first-time-setup.md: Onboards by creating dirs, linking Notion/Obsidian via MCP.
💐 greenhouse.md: "Show me the greenhouse" dashboards vitals/themes/ripeness/convergences/orphans.
🌱 plant.md: "Plant this" sorts input as new seed or signal (checks garden-state.md clusters first; asks germination if URL).
🌾 ripen.md: "Ripen" audits 2+ criteria seeds, lists missing steps, auto-moves at threshold.
🍂 compost.md: "Compost" flags wilting (14 days no activity) or orphans (10 days no connections), cross-refs compost for revivals.

Architecture scales: index-first reading keeps it fast at 100+ seeds; 98-end-of-session.md updates state/memory for persistence.

Germination and 3/5 Ripeness Threshold Yield Writable Stakes

Capture personal stake at planting with 2 questions: "What made you notice this?" (your perspective) and "Do you agree or resist it?" (tension). Seeds format as MD files with ID, status, dates, signals count, ripeness score, sections (original signal, germination, attached signals, agent notes, history).

Ripen at 3/5 criteria:

Signal Diversity (2+ source types; avoids echoes).
Cluster Size (2+ seed connections).
Tension Present (unresolved question/contrarian angle; turns observation into content).
Personal Stake (your engagement beyond planting).
Age Threshold (14+ days; rewards slow thinking).

Composting logic: auto-surfaces candidates with context; revives via new connections. Kills cognitive load—single "Plant" entry (agent sorts), immediate germination (no queues), literal file-as-plant (no abstract clusters).

Impact: Newsletter writers connect weekly fragments over 3 weeks; consultants spot client patterns after 4 weeks; pros surface 14 angles from expertise.

Deploy in Under 5 Minutes or Build from Spec

Premium users download from RobotsOS, point Claude/Codex to folder, run "Help me onboard the greenhouse agent." Free: Paste article into Claude with build prompt extracting full spec (dirs, rules, skills, criteria). Ask agent for workflows (e.g., "Suggest best rhythm"), commands list, or explanations (e.g., "What is germination?").