docs: reorder README to lead with what OpenWar does

Three changes per operator feedback:

1. Drop "What OpenWar is not" from the opening section. The first H2
   was framing the project negatively before saying what it does. Now
   the first H2 is "Just talk to it" - positive, action-oriented.

2. Move "What OpenWar is not" later, after "Why both a framework AND
   a runtime". Clarifications belong after the positive framing is in.
   Also refreshed the entries: dropped the stale v0.6 placeholder
   (already shipped) and added "not a self-modifying agent" which is
   the architectural NO that distinguishes operator-gated learning
   from agent-gated self-modification.

3. Collapse the 13-bullet version history. CHANGELOG.md is the source
   of truth for release notes; the README just points there. Current
   version bumped to v0.11.0.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: PythonLuvr

README.md

@@ -26,17 +26,7 @@ Discipline, not intelligence. Kubernetes for agents, not a smarter brain.
   <em>Default agent behavior. Sycophantic, eager, drowning in half-finished context.</em>
 </p>
 
-## What OpenWar is not
-
-Not a smarter model. OpenWar runs on top of whatever agent you already use: a local CLI agent (Claude Code, Codex CLI, Gemini CLI, aider) via the cli-bridge adapter, or any BYOK API (Anthropic, OpenAI, Gemini, Grok, or openai-compat for local Ollama and vLLM). The model's reasoning quality is the model's problem.
-
-Not a memory system or knowledge base. Persistent project memory is planned for v0.6, but the current scope is behavioral discipline, not cognitive infrastructure.
-
-Not an autonomous-agent platform. OpenWar's whole design assumes the operator is in the loop. Auto-pilot mode just makes the loop quieter; it never removes the operator's right to stop the run.
-
----
-
-## New in v0.10: just talk to it
+## Just talk to it
 
 If you have a BYOK API key (Anthropic, OpenAI, Gemini, Grok, or any OpenAI-compatible endpoint):
 
@@ -160,23 +150,17 @@ System prompts cost nothing to install and work with any agent. The runtime is h
   <em>Phase 4 completion. WarBit ships.</em>
 </p>
 
+## What OpenWar is not
+
+Not a smarter model. OpenWar runs on top of whatever agent you already use: a local CLI agent (Claude Code, Codex CLI, Gemini CLI, aider) via the cli-bridge adapter, or any BYOK API (Anthropic, OpenAI, Gemini, Grok, or openai-compat for local Ollama and vLLM). The model's reasoning quality is the model's problem.
+
+Not a self-modifying agent. The framework can propose adjustments to its own behavior (`openwar learn`), but the operator commits the change. The agent never edits the rules that constrain the agent.
+
+Not an autonomous-agent platform. OpenWar's whole design assumes the operator is in the loop. Auto-pilot mode just makes the loop quieter; it never removes the operator's right to stop the run.
+
 ## Versioning
 
-Current: **v0.10.0**. See [CHANGELOG.md](./CHANGELOG.md) for full release notes.
-
-- v0.1: framework doc only (single markdown file).
-- v0.2: runtime, CLI, BYOK adapters.
-- v0.3: native tools + MCP client + Phase 3 destructive flag for unauthorized tool calls.
-- v0.4: multi-agent orchestration. Planner / executor / reviewer / critic. Budgets. Resumable mid-state sessions.
-- v0.5: cli-bridge adapter. OpenWar coordinates CLI agents (Claude Code, Codex CLI, Gemini CLI, aider) the same way it coordinates LLM adapters.
-- v0.5.1: per-role adapter mixing. A single brief can pin each role to its own adapter and model.
-- v0.6: persistent project memory across briefs.
-- v0.7: cli-bridge MCP-server-mode (Claude Code, Gemini CLI, Codex CLI) + Claude Code permission auto-setup + symmetric memory tools.
-- v0.8: structured trace event stream, focused `openwar inspect` modes, `openwar replay` (no LLM calls), opt-in local dashboard.
-- v0.9.0: `openwar history` (descriptive analytics over accumulated v0.8 traces). Read-only.
-- v0.9.1: adaptive autonomy plumbing. `openwar learn`, `learned_profile:` frontmatter, detector sensitivity, three new trace events. Conservative thresholds: effectively a no-op until run ~10. v0.9.2+ patch releases tune the constants.
-- v0.10.0: `openwar chat`. Conversation-driven entry to the same runtime. Tool-call intent contract; conservative auto-authorization; saved-brief round trip.
-- v0.10.1 (planned): hero rewrite + mid-tool-call cancellation, pending adoption signal from v0.10.0 use.
+Current: **v0.11.0**. Full version history and release notes in [CHANGELOG.md](./CHANGELOG.md).
 
 Drop-in upgrades preserve compatibility within a major version. Major bumps may rename phases or change the brief format.