Add an AGENTS.md pointing to our policy (#156838)

Co-authored-by: 🇺🇦 Sviatoslav Sydorenko (Святослав Сидоренко) <wk.cvs.github@sydorenko.org.ua>
Co-authored-by: Łukasz Langa <lukasz@langa.pl>
Co-authored-by: Lukasz Langa <ambv@meta.com>
Co-authored-by: Muse Code (Meta Muse Spark)

.claude/CLAUDE.md

@@ -0,0 +1 @@
+../AGENTS.md

.gitignore

@@ -178,8 +178,12 @@ Python/frozen_modules/MANIFEST
 /python
 !/Python/
 
-# People's custom https://docs.anthropic.com/en/docs/claude-code/memory configs.
-/.claude/
+# Local AI agent scratch state (per-PR and per-branch notebooks, sandbox
+# experiments) and personal agent overrides, none of which are committed.
+/.claude/pr-*
+/.claude/branch-*
+/.claude/sandbox/
+AGENTS.local.md
 CLAUDE.local.md
 
 #### main branch only stuff below this line, things to backport go above. ####

AGENTS.md

@@ -0,0 +1,16 @@
+# AI agent guidance
+
+CPython has a [policy on the use of AI tools](https://devguide.python.org/getting-started/ai-tools/).
+All use of AI tools and agents when working on or interacting with CPython
+must follow it.
+
+> [!important]
+> **Primary directive**: Read the policy before making or proposing any changes.
+
+When acting on this repository, apply the policy's core principles:
+
+- Consider whether the change is necessary.
+- Make minimal, focused changes.
+- Follow existing coding style and patterns.
+- Write tests that exercise the change.
+- Keep backwards compatibility with prior releases in mind.