chore: baseline converge (CI fix + agent stack)

[antongulin]

Automated baseline converge pass (2026-07-11):

• actions: read on the robin caller (required since robin#70; fixes fleet-wide startup_failure)
• robin LLM secrets rotated to OpenRouter (done directly, not in this diff)
• DOX block + engineering-baseline pointer appended to AGENTS.md (idempotent, non-destructive)
• CLAUDE.md/GEMINI.md symlinks where absent (pre-existing real files left untouched, warned)

Law: ~/Code/700 systems/baseline. bootstrap rc=0

🤖 Generated with Claude Code

[github-actions]

🏹 Robin

✅ Review done. I flagged 3 things worth a look.

Push fixes whenever you like, then comment /robin for another pass.

[github-actions]

🏹 Robin

Heads up: this is a point-in-time review. Push fixes freely, then comment /robin whenever you want another pass.

🚨 1 High | ⚠️ 1 Medium | 💡 1 Suggestions

Summary

The change adds an AGENTS.md framework file with symlinks from CLAUDE.md and GEMINI.md, and grants an additional actions: read permission in the code-review workflow. The workflow permission addition is reasonable and well-commented. However, the AGENTS.md file contains an embedded active instruction designed to make AI agents scan and rewrite the repository, which is a prompt-injection risk, and references a developer-specific local filesystem path that will not resolve for other contributors.

---

Robin — the Robin Hood of code review. Free for every PR.