Clarify Rally value prop and architecture ownership in docs

[justinmoon]

Summary\n- rewrite README intro around Rally's one-instruction-at-a-time coordinator model\n- document current build-time workflow extension model and runtime plugin roadmap status\n- add AGENTS.md with concise ownership boundaries for core vs workflow crates vs wrapper prompts\n\n## Why\n- make the core value proposition and intended architecture clearer\n- reduce prompt/workflow ownership drift

---

Summary by CodeRabbit

Documentation

• Clarified project description and core purpose as a turn-based coordinator
• Expanded architecture boundaries documentation to distinguish core responsibilities, workflow crates, and wrapper-specific content
• Updated command examples and terminology to reflect current usage patterns
• Added agent notes documentation for workflow and wrapper guidance

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: fc1bcc0d-13c0-48a3-ae0f-54e5642002b5

📥 Commits

Reviewing files that changed from the base of the PR and between 8613ffa and 912053c.

📒 Files selected for processing (2)

• AGENTS.md
• README.md

---

📝 Walkthrough

Walkthrough

Rally's documentation is refactored to emphasize its role as a "turn-based coordinator" rather than a "step-by-step coordinator." A new AGENTS.md file is added to document agent-specific notes, architecture ownership, and content maintenance rules. README.md is substantially rewritten to clarify architectural boundaries, core responsibilities, workflow-crate ownership, and shift focus from wrapper-centric usage to system architecture rationale.

Changes

Cohort / File(s)

Summary

Documentation Updates AGENTS.md, README.md

New AGENTS.md file added with agent notes, architecture ownership guidelines, and content maintenance rules. README.md rewritten to reframe Rally as a turn-based coordinator, emphasizing architectural boundaries (core vs. workflow crates), shifting from wrapper-centric to architecture-focused narrative, updating canonical entry points (rally skill run/exec), and replacing runtime workflow examples with reference documentation.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Poem

🐰 Hops through docs with careful paws,
New architecture without flaws,
Rally's heart now crystal clear—
Turn-based wisdom, far and near!
📚✨

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch docs/core-value-prop

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[devin-ai-integration]

Devin Review found 1 potential issue.

View 2 additional findings in Devin Review.

[devin-ai-integration]

🔴 AGENTS.md rule violation: rally skill context added to README.md without updating docs/skill-install-run.md

The PR adds rally skill context to the README's wrapper-helpers list (line 56), but docs/skill-install-run.md is not updated in the same change. This violates the AGENTS.md practical rule: "If you add/rename a run command, update wrapper-facing docs (README.md, docs/skill-install-run.md) in the same change." The old README (git show HEAD~1:README.md) did not list rally skill context, and docs/skill-install-run.md has zero mentions of skill context — confirmed via grep.

Prompt for agents

Add documentation for the `rally skill context` command to docs/skill-install-run.md to satisfy the AGENTS.md rule that requires both README.md and docs/skill-install-run.md to be updated in the same change when a command is added. The context command is already implemented in src/command_surface.rs:194 and src/cli.rs:352-355. Add a section to docs/skill-install-run.md (e.g. after the existing `rally skill` Patterns section around line 63) documenting the `rally skill context` subcommand, its purpose (printing workspace context for intelligent command resolution), and its arguments (see SkillContextArgs in src/cli.rs:410).

---

Was this helpful? React with 👍 or 👎 to provide feedback.