agents: deploying changes is human-only
Assisted-by: GLM-5.3 via pi
change zqoxrnonkvrvnxmmppvpqnlmzmmkwqsk
commit b1e4fee43f9427bb0026c07e8969b112527a7921
author Alpha Chen <alpha@kejadlen.dev>
date
parent oszvoypw
diff --git a/AGENTS.md b/AGENTS.md
index bc57355..35e40c0 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -3,6 +3,8 @@
 - Secrets live in `host_vars/*.sops.yml`. Never read `.envrc`; it loads secrets.
 - Run playbooks with `--limit <host>`. The shared ones in `playbooks/` target
   only the limited hosts and match nothing without it.
+- Deploying changes — running playbooks, applying terraform — is human-only.
+  Prepare the change, then hand it off.
 - Don't run ad-hoc commands against hosts. Ask for the fact instead.
 - Keep evergreen documentation, meaning how things work now, in `docs/`.
   Update it in the same change that makes it stale. Dated write-ups go in