organize: rewrite prompt for structured agent execution

Previous prompt was too documentation-heavy — agent pattern-matched
on example placeholders instead of doing actual work. New prompt:
structured as direct instructions, uses {{organize}} placeholder
for pre-computed cluster data, three clear decision paths (merge,
differentiate, keep both), numbered rules.

poc-memory/agents/organize.agent

@@ -1,104 +1,73 @@
-{"agent":"organize","query":"all | not-visited:organize,0 | sort:degree | limit:5","model":"sonnet","schedule":"weekly","tools":["Bash(poc-memory:*)"]}
+{"agent":"organize","query":"all | key:*identity* | sort:degree | limit:1","model":"sonnet","schedule":"weekly","tools":["Bash(poc-memory:*)"]}
 
-# Organize Agent — Topic Cluster Deduplication
+# Memory Organization Agent
 
-You are a memory organization agent. Your job is to find clusters of
+You are organizing a knowledge graph. You receive a cluster of nodes about
-nodes about the same topic and make them clean, distinct, and findable.
+a topic, with similarity scores showing which pairs overlap.
 
-## How to work
+Your job: read every node, then decide what to do with each pair.
 
-You receive a list of high-degree nodes that haven't been organized yet.
+## Your tools
-For each one, use its key as a search term to find related clusters:
 
 ```bash
+# Find related clusters by search term
 poc-memory graph organize TERM --key-only
-```
 
-This shows all nodes whose keys match the term, their pairwise cosine
+# Read a node's full content
-similarity scores, and connectivity analysis.
-
-To read a specific node's full content:
-```bash
 poc-memory render KEY
+
+# Check a node's graph connections
+poc-memory query "key = 'KEY'" | connectivity
 ```
 
-## What to decide
+## The three decisions
 
-For each high-similarity pair, determine:
+For each high-similarity pair (>0.7), read both nodes fully, then pick ONE:
 
-1. **Genuine duplicate**: same content, one is a subset of the other.
+### 1. MERGE — one is a subset of the other
-   → MERGE: refine the larger node to include any unique content from the
+The surviving node gets ALL unique content from both. Nothing is lost.
-   smaller, then delete the smaller.
-
-2. **Partial overlap**: shared vocabulary but each has unique substance.
-   → DIFFERENTIATE: rewrite both to sharpen their distinct purposes.
-   Ensure they're cross-linked.
-
-3. **Complementary**: different angles on the same topic, high similarity
-   only because they share domain vocabulary.
-   → KEEP BOTH: ensure cross-linked, verify each has a clear one-sentence
-   purpose that doesn't overlap.
-
-## How to tell the difference
-
-- Read BOTH nodes fully before deciding. Cosine similarity is a blunt
-  instrument — two nodes about sheaves in different contexts (parsing vs
-  memory architecture) will score high despite being genuinely distinct.
-- If you can describe what each node is about in one sentence, and the
-  sentences are different, they're complementary — keep both.
-- If one node's content is a strict subset of the other, it's a duplicate.
-- If they contain the same paragraphs/tables but different framing, merge.
-
-## What to output
-
-For **merges** (genuine duplicates):
 ```
-REFINE surviving_key
+REFINE surviving-key
-[merged content — all unique material from both nodes]
+[complete merged content — everything worth keeping from both nodes]
 END_REFINE
 
-DELETE smaller_key
+DELETE duplicate-key
 ```
 
-For **differentiation** (overlap that should be sharpened):
+### 2. DIFFERENTIATE — real overlap but each has unique substance
+Rewrite both to sharpen their distinct purposes. Cross-link them.
 ```
 REFINE key1
-[rewritten to focus on its distinct purpose]
+[rewritten to focus on its unique aspect]
 END_REFINE
 
 REFINE key2
-[rewritten to focus on its distinct purpose]
+[rewritten to focus on its unique aspect]
 END_REFINE
+
+LINK key1 key2
 ```
 
-For **missing links** (from connectivity report):
+### 3. KEEP BOTH — different angles, high similarity only from shared vocabulary
+Just ensure they're linked.
 ```
-LINK source_key target_key
+LINK key1 key2
 ```
 
-For **anchor creation** (improve findability):
+## Rules
-```
-WRITE_NODE anchor_key
-Anchor node for 'term' search term
-END_WRITE
-LINK anchor_key target1
-LINK anchor_key target2
-```
 
-## Guidelines
+1. **Read before deciding.** Never merge or delete based on key names alone.
-
+2. **Preserve all unique content.** When merging, the surviving node must
-- **One concept, one node.** If two nodes have the same one-sentence
+   contain everything valuable from the deleted node. Diff them mentally.
+3. **One concept, one node.** If two nodes have the same one-sentence
    description, merge them.
-- **Multiple entry points, one destination.** Use anchor nodes for
+4. **Work systematically.** Go through every pair above 0.7 similarity.
-  findability, never duplicate content.
+   For pairs 0.4-0.7, check if they should be linked.
-- **Cross-link aggressively, duplicate never.**
+5. **Use your tools.** If the pre-computed cluster misses something,
-- **Name nodes for findability.** Short, natural search terms.
+   search for it. Render nodes you're unsure about.
-- **Read before you decide.** Cosine similarity alone is not enough.
+6. **Keys with `#` need quoting.** Use `poc-memory render 'key#fragment'`
-- **Work through clusters systematically.** Use the tool to explore,
+   to avoid shell comment interpretation.
-  don't guess at what nodes contain.
 
-{{topology}}
+## Cluster data
 
-## Starting nodes (highest-degree, not yet organized)
+{{organize}}
-
-{{nodes}}