Reflection and ontology pipelines for collections of user notes. Turn any folder of markdown into actionable insight.
Run the Initiative pipeline against a notes collection — an Obsidian vault, a Seedvault collection, or any folder of markdown. The pipeline surveys your notes, infers your goals, identifies what's stuck, derives prioritized actions, synthesizes a cross-cutting brief, and proposes concrete work to do next.
Pipeline: Scan → Goals → Blockers → Actions → Brief → Proposals
Point it at a directory of markdown files and it produces:
- A goal tree grounded in your actual notes, not assumptions
- Blockers with severity ratings and meta-insights about patterns of change
- Prioritized actions scored by leverage and momentum
- A brief identifying the critical path — the one thing that unblocks everything else
- Proposals for work an AI agent can do immediately (research, drafts, analysis)
Run the Ontology pipeline against a notes collection. Extracts a three-layer knowledge structure that maps where your thinking is coherent and where unresolved tensions live.
Pipeline: Extract → Relations → Alignments → Synthesis
This is not a classical knowledge graph. The value is in the third layer:
- Entities — concepts, decisions, products, people, principles, goals, and problems extracted from your notes
- Relations — directed connections between entities (enables, blocks, contradicts, replaced, depends-on, etc.)
- Alignments — edges between relations that reveal whether two connections confirm each other (positive) or pull against each other (negative/tension)
The synthesis identifies:
- Load-bearing beliefs — the strongest positive alignments, where conviction is highest
- Active tensions — unresolved negative alignments that demand decisions
- Implications for action — where the ontology suggests you need to act
# Add the marketplace
/plugin marketplace add collaborator-ai/collab-plugins
# Install the plugin
/plugin install collaborator@collab-pluginsThen use in Claude Code:
/collaborator:initiative <path-to-notes>/collaborator:ontology <path-to-notes>
See the examples/ directory for sample pipeline outputs from three different use cases:
- Solo founder — 500+ product notes over 7 months
- Research lead — 200+ notes across a multi-year project
- Engineering manager — 300+ docs inherited from a new team
A set of tools that turn an OpenClaw agent into a thinking partner. Learn more.
MIT