From ea20cf73432d91c4371817a21ad7092c3ea687b4 Mon Sep 17 00:00:00 2001
From: "detail-app[bot]" <180357370+detail-app[bot]@users.noreply.github.com>
Date: Thu, 30 Jul 2026 15:58:40 +0000
Subject: [PATCH] docs: correct LM1c to accept any non-blank gateway chat alias

---
 AGENTS.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/AGENTS.md b/AGENTS.md
index 92acd652..fd76383c 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -223,7 +223,7 @@ alwaysApply: true
 
 - [LM1a] **Settings**: Do not change any LLM settings without explicit written approval.
 - [LM1b] **No Fallback**: Do not auto-fallback or regress models across providers; surface error to user.
-- [LM1c] **Config**: Generative inference MUST use exact `gpt-5.4` through the shared OpenAI-compatible gateway. Its model/base-url/api-key MUST come from `.env`/environment variables (see [EV1d]). All other LLM settings MUST come from Spring property files and `@ConfigurationProperties`.
+- [LM1c] **Config**: Generative inference MUST go through the shared OpenAI-compatible gateway using a non-blank `OPENAI_MODEL` gateway alias (default `gpt-5.4`). Its model/base-url/api-key MUST come from `.env`/environment variables (see [EV1d]). All other LLM settings MUST come from Spring property files and `@ConfigurationProperties`.
 - [LM1d] **Behavior**: Allowed: logging diagnostics. Not allowed: silently changing LLM behavior.
 - [LM1e] **Streaming**: TTFB < 200ms, streaming start < 500ms.
 - [LM1f] **Events**: `text`, `citation`, `code`, `enrichment`, `suggestion`, `status`.