Add LinaAI automatic Codex routing

Docs/AI/LocalAgentGuardrails.md

@@ -19,7 +19,7 @@ These rules apply to any self-hosted AI coding assistant working on Agrarian.
 
 Stop local work and prepare a Codex handoff when any of these are true:
 
-- confidence is below `0.65`,
+- confidence is below `0.75`,
 - tests fail twice,
 - build fails twice,
 - Unreal compile errors persist after one focused fix,

Docs/AI/SelfHostedAiDevelopmentStack.md

@@ -30,13 +30,15 @@ Codex escalation when local tooling is over its head.
 
 ## Operating Model
 
-1. Local AI gathers context and proposes small changes.
-2. Work happens on a branch, not directly on `main`.
-3. The agent reports risk, files inspected, commands run, and confidence.
-4. Tests/builds decide whether a change is acceptable.
-5. After two failed local attempts, stop and escalate.
-6. Codex escalation uses the npm Codex CLI, not the API.
-7. Human review controls merges.
+1. Start with `Scripts/linaai_task.sh`, not raw Aider, for normal work.
+2. Qwen/Ollama performs a preflight risk and confidence check.
+3. Default confidence threshold is `0.75`.
+4. High-risk tasks or low-confidence tasks route to Codex automatically.
+5. Aider runs only for acceptable supervised local work.
+6. If Aider fails, `Scripts/linaai_task.sh` writes a status file and calls
+   Codex through `Scripts/ai_codex_escalate.sh`.
+7. Codex escalation uses the npm Codex CLI, not the API.
+8. Human review controls merges.
 
 ## Codex Escalation
 
@@ -44,6 +46,19 @@ Use `Scripts/ai_codex_escalate.sh` with a completed task status file. The
 script prefers a locally installed `codex` command and falls back to
 `npx -y @openai/codex exec`.
 
+For normal tasks, use:
+
+```bash
+cd ~/repos/AgrarianGame
+Scripts/linaai_task.sh "your task here"
+```
+
+To test automatic escalation without editing files:
+
+```bash
+Scripts/linaai_task.sh --dry-run --force-escalate "Test escalation path only."
+```
+
 On `LinaAI`, the npm Codex CLI is installed, but it still needs an authenticated
 Codex login before cloud escalation can run:
 
@@ -54,7 +69,7 @@ codex login
 
 Codex should be called for:
 
-- confidence below `0.65`,
+- confidence below `0.75`,
 - two failed build/test attempts,
 - Unreal compile errors that persist,
 - tasks touching save systems, multiplayer, auth, payments, AGR wallet

Docs/Ops/HANDOFF.md

@@ -57,6 +57,24 @@
 - `codex doctor` on `LinaAI` reports the npm Codex CLI install is healthy but
   not authenticated yet. Run `codex login` on `LinaAI` before expecting Codex
   escalation to execute.
+
+## LinaAI Automatic Aider To Codex Routing - 2026-05-24
+
+- Added `Scripts/linaai_task.sh` as the normal LinaAI task entry point.
+- Default local confidence threshold is `0.75`; `0.65` is too permissive for
+  the current project risk profile.
+- Workflow:
+  - Qwen/Ollama performs preflight risk and confidence classification.
+  - high-risk tasks route directly to Codex.
+  - tasks below `0.75` confidence route directly to Codex.
+  - acceptable tasks run through Aider with `--no-auto-commits`.
+  - if Aider exits unsuccessfully, the script writes a status file and calls
+    `Scripts/ai_codex_escalate.sh`.
+- High-risk keyword routing includes Unreal core architecture, save/load,
+  multiplayer, networking/replication, AGR wallet/payments, marketplace/economy
+  transfer logic, auth, security, migrations, secrets, and broad refactors.
+- Test command:
+  `Scripts/linaai_task.sh --dry-run --force-escalate "Test escalation path only."`
 - Added self-hosted AI project documentation:
   - `Docs/AI/SelfHostedAiDevelopmentStack.md`
   - `Docs/AI/LocalAgentGuardrails.md`