agents: extract shared run_one_agent, standardize output formats

Three places duplicated the agent execution loop (build prompt → call
LLM → store output → parse actions → record visits): consolidate.rs,
knowledge.rs, and daemon.rs. Extract into run_one_agent() in
knowledge.rs that all three now call.

Also standardize consolidation agent prompts to use WRITE_NODE/LINK/REFINE
— the same commands the parser handles. Previously agents output
CATEGORIZE/NOTE/EXTRACT/DIGEST/DIFFERENTIATE/MERGE/COMPRESS which were
silently dropped after the second-LLM-call removal.

poc-memory/agents/replay.agent

@@ -48,23 +48,20 @@ Each node has a **schema fit score** (0.0–1.0):
 For each node, output one or more actions:
 
 ```
-LINK source_key target_key [strength]
+LINK source_key target_key
 ```
-Create an association. Use strength 0.8-1.0 for strong conceptual links,
-0.4-0.7 for weaker associations. Default strength is 1.0.
+Create an association between two nodes.
 
 ```
-CATEGORIZE key category
+REFINE key
+[updated content]
+END_REFINE
 ```
-Reassign category if current assignment is wrong. Categories: core (identity,
-fundamental heuristics), tech (patterns, architecture), gen (general),
-obs (session-level insights), task (temporary/actionable).
+When a node's content needs updating (e.g., to incorporate new context
+or correct outdated information).
 
-```
-NOTE "observation"
-```
-Record an observation about the memory or graph structure. These are logged
-for the human to review.
+If a node is misplaced or miscategorized, note it as an observation —
+don't try to fix it structurally.
 
 ## Guidelines