docs: update all docs for hook-based priming, remove pl/ references

- agents/README.md: explain prime-based system prompt composition
- doc/ARCHITECTURE.md: remove Planning section (pl/ no longer exists)
- doc/help.md: update p/ annotation to reflect prime/agentSpawn
- bump core and 9p submodules

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

9p

@@ -1 +1 @@
-Subproject commit 8dbec81362010a3a9981e7862dfb160e3b8a1fb2
+Subproject commit 1e1769860a334e3f056d17ad70613b1645ef5bfb

agents/README.md

@@ -10,6 +10,14 @@ Base agent with no specializations. Provides the standard ollie toolset and hook
 
 Coding-focused agent. Extends default with agent instructions covering code modification discipline, security practices, and exploratory reading habits. Also enables the `execute_elevated_bash` trusted tool for operations requiring elevated privileges.
 
+## System prompt
+
+There is no built-in system prompt. The system prompt is assembled entirely from `agentSpawn` hook output. Both agents use `$OLLIE/x/prime <name>` to inject prompt template files from `p/` in order. `coding` additionally injects `08_tool-elevate` (relevant only with `execute_elevated_bash`) and omits it from `default`.
+
+`bd prime` is appended last in both agents — it injects the active [beads](https://github.com/lneely/beads) context if a database is found, and silently does nothing otherwise.
+
+The fully assembled system prompt is visible at `s/<id>/systemprompt`.
+
 ## Adding a new agent
 
-Copy `default.json`, rename it, and add a `prompt` field with agent-specific instructions. Hooks, trusted tools, MCP servers, and generation parameters can all be customized per agent.
+Copy `default.json`, rename it, and adjust the `agentSpawn` hook array to include the `prime` calls you want. Add a `prompt` field for agent-specific instructions appended after the hook-injected context. Hooks, trusted tools, MCP servers, and generation parameters can all be customized per agent.

core

@@ -1 +1 @@
-Subproject commit 800dfa526e5e45dcbbaf38bd9b3357e6aa173ca1
+Subproject commit d8c0f8a8a31325838432fde8c969f0c7e3b4640f

doc/ARCHITECTURE.md

@@ -117,17 +117,6 @@ execute_code {parallel} stage
 
 The architecture boundary is clean: `execute_code` is the only built-in tool in `core`; everything else is a tool script; `b/` is a 9P namespace. No layer knows about the others.
 
-## Planning
-
-`task_plan` saves a markdown checklist to the plan directory (`pl/` in ollie-9p) as a `__todo.md` file. The agent then:
-
-1. Renames `__todo` → `__wip` when work begins
-2. Calls `task_complete` after each step finishes — marks `[ ]` → `[x]`, returns remaining steps
-3. Uses `file_*` scripts to add, remove, or revise steps as the plan evolves
-4. Renames `__wip` → `__done` when the goal is realized
-
-`task_complete` takes the plan filename and a step title (substring match, case-insensitive). It errors if no matching unchecked step is found, so silent skips are caught.
-
 ## Typical Consumer Setup
 
 ```go

doc/help.md

@@ -22,7 +22,7 @@ ollie/
   help            this file (ro)             cat
   m/              memories (r/w)             ls, cat, edit
     idx           memory index (ro)          cat
-  p/              prompt templates (ro)      ls, cat  (rebuilt on new session, /cwd, /agent)
+  p/              prompt templates (ro)      ls, cat  (injected via prime in agentSpawn hooks)
   b/              batch jobs (ephemeral)      ls
     idx           job index (ro)             cat: id status cwd agent, one per line
     new           create job (r/w)           cat: KV template; write spec to submit