Add AGENTS.md and CLAUDE.md for AI agent guidance

[rtibbles]

Summary

Add two documentation files to guide AI coding agents working in this repository:

• AGENTS.md — Concise, model-agnostic guidance (36 lines) with commands-first layout, critical gotchas, project overview, and key architecture. Designed to fit in smaller context windows.
• CLAUDE.md — Imports AGENTS.md via @AGENTS.md reference and adds Claude Code-specific extensions: detailed pipeline architecture, le_utils dependency context, and expanded file type implementation guidance.

The AGENTS.md content was validated through 3 rounds of TDD testing (RED-GREEN-REFACTOR) using Haiku and Sonnet subagents with a KPUB file type implementation task. Each round identified common agent mistakes and refined the gotchas to prevent them. Key gotchas that proved effective:

• Preventing agents from creating new File/Node subclasses in legacy classes/ modules
• Requiring le_utils constant verification before use (names are non-obvious)
• Enforcing correct test placement in tests/pipeline/
• Preventing over-copying of validation logic from existing handlers

References

Follows the AGENTS.md / CLAUDE.md split pattern from Kolibri PR #14164.

Reviewer guidance

• These are documentation-only changes (no code modifications)
• AGENTS.md Critical Gotchas section — Does this accurately capture the most common agent mistakes when working in this codebase? Are there other gotchas that should be added?
• CLAUDE.md extended architecture — Does this provide enough context for Claude Code without being redundant with AGENTS.md?
• The TDD validation found that HTML body parsing over-copying and File subclass creation (for Haiku) remain hard to prevent through documentation alone — these may need structural guardrails in the future