Distillary

Agent Retrieval Skill

2 min read

Agent retrieval skill

A self-contained skill for querying any published Distillary brain. Copy this into any agent’s context — it only needs a brain URL and HTTP GET.

The skill

Copy the contents of .claude/skills/distillary-retrieval.md into your agent’s system prompt or skill library.

How it works

  1. Agent fetches {brain_url}/agent.json — lightweight manifest with sources, bridges, and navigation instructions
  2. Based on the question type, the agent picks a strategy:
    • Concept lookup → fetch a bridge entity page → backlinks ARE the answer
    • Source summary → thesis is already in the manifest
    • Comparison → fetch the comparison essay or a bridge page
    • Deep evidence → walk the pyramid: root → structure → cluster → atom
    • Exploration → follow backlink chains through the concept graph
  3. Each fetch returns one page with content + links to go deeper
  4. Most questions answered in 2-4 fetches, under 2500 tokens

Example

Agent receives question: “What does this brain say about validating demand?”

GET /agent.json
→ Bridge: "Demand Test" — testing whether customers want what you're building

GET /shared/concepts/demand-test
→ Lean Startup: ship an MVP and measure real behavior
→ Mom Test: check for commitment that costs something (time, money, reputation)
→ Both agree: compliments don't count as validation

Done. 2 fetches. Multi-source answer.

Integration

Works with any agent that can make HTTP requests:

  • Claude Code (WebFetch tool)
  • ChatGPT (web browsing)
  • Custom agents (curl/fetch)

No API keys. No authentication. No MCP server. Just URLs and structured markdown pages.

Graph View

Backlinks