Skip to content

docs(agents): condense AGENTS.md 196->40 lines#998

Merged
buremba merged 1 commit into
mainfrom
chore/slim-agents
May 21, 2026
Merged

docs(agents): condense AGENTS.md 196->40 lines#998
buremba merged 1 commit into
mainfrom
chore/slim-agents

Conversation

@buremba
Copy link
Copy Markdown
Member

@buremba buremba commented May 21, 2026

AGENTS.md inlines into every session (and re-injects per worktree), so its length is a per-task token cost. Condenses verbose prose + pushes deep subsystem reference (guardrails internals, egress/network detail, validation recipes) to see <code path> pointers.

All rules preserved (verified): worktree-only writes, multi-replica/ClientIP correctness gate, secret-proxy, events append-only, workspace:* deps, dynamic-import allow-list, no git stash, isolation worktrees, Node 22-24, make review + E2E gate, submodule push-before-bump + relock, DESIGN_GUIDELINES read-on-demand. −198/+40.

Summary by CodeRabbit

  • Documentation
    • Updated internal operating procedures and development guidelines for clarity and accuracy.

Note: This release contains no user-facing feature changes or updates.

Review Change Stack

AGENTS.md is inlined into every session via CLAUDE.md and re-injected per
worktree, so its length is paid on every task. Condense verbose prose and
push deep subsystem reference (guardrails internals, egress/network detail,
validation recipes) to "see <code path>" pointers.

All rules preserved: worktree-only writes, the multi-replica/ClientIP
correctness gate, secret-proxy, events append-only, workspace:* deps,
the dynamic-import allow-list, no git stash, isolation worktrees, Node 22-24,
make review + E2E gate, submodule push-before-bump + relock, DESIGN_GUIDELINES
read-on-demand. Reference detail moved to code, not deleted.
@buremba buremba merged commit a2db4c6 into main May 21, 2026
@buremba buremba deleted the chore/slim-agents branch May 21, 2026 00:24
@coderabbitai
Copy link
Copy Markdown

coderabbitai Bot commented May 21, 2026

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 2482f4f2-d067-459c-8ebe-5b27baa93bf9

📥 Commits

Reviewing files that changed from the base of the PR and between 3ec1937 and 4dc043f.

📒 Files selected for processing (1)
  • AGENTS.md

📝 Walkthrough

Walkthrough

AGENTS.md is reorganized into six clearer operational sections: Structure (workspace and submodule setup), Architecture (platform isolation and multi-replica constraints), Releasing (release-please ownership and CI flow), Agent rules (worktree workflows), Branch hygiene (commit constraints), and Development & validation (testing gates and secrets). Content tightens platform isolation language, multi-replica correctness rules, connector dependency consolidation, guardrail configuration, and reaffirms worker-session persistence.

Changes

Operating Instructions Reorganization

Layer / File(s) Summary
Foundational platform and architecture rules
AGENTS.md
Workspace/Bun preferences, submodule push-and-lockfile instructions, frontend design guidelines, tightened platform isolation and multi-replica/SSE delivery constraints, orchestration/worker subprocess expectations, MCP and connector dependency/build pipeline rules, guardrail configuration and trip behaviors, and network egress/judge constraints.
Release, workflow, and validation rules
AGENTS.md
Release-please ownership, workspace:* dependencies, CI publishing flow, worktree-first editing and make command directives, Slack link handling, planning/code-import allow-list entries, branch/commit hygiene constraints (no stash, worktree isolation), validation and bot/evals guidance, .env as secrets source, and reaffirmed worker-session persistence.

Possibly related PRs

  • lobu-ai/lobu#962: Both PRs are documentation-only updates to AGENTS.md that strengthen multi-replica (N>1) Kubernetes correctness rules and pod-local in-memory state expectations with emphasis on cross-replica delivery via Postgres.
  • lobu-ai/lobu#941: Both PRs rewrite AGENTS.md to reorganize agent rules around platform isolation, worker/orchestration constraints, guardrails, egress/judge constraints, and validation workflow in the same logical areas.
  • lobu-ai/lobu#928: Both PRs add or adjust worktree-only editing and submodule bump workflow guidance to AGENTS.md documentation.

Poem

🐰 The agents now know their path so clear,
Platforms isolated, replicas cheer!
From structure to release, each rule finds its place,
Worktrees and workflows in neat, ordered grace. ✨

🎯 2 (Simple) | ⏱️ ~10 minutes

✨ Finishing Touches
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Commit unit tests in branch chore/slim-agents

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

@codecov-commenter
Copy link
Copy Markdown

⚠️ Please install the 'codecov app svg image' to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants