5chan/AGENTS.md

AGENTS.md

Purpose

This file defines the always-on rules for AI agents working on 5chan. Use this as the default policy. Load linked playbooks only when their trigger condition applies.

Surprise Handling

The role of this file is to reduce recurring agent mistakes and confusion points in this repository. If you encounter something surprising or ambiguous while working, alert the developer immediately. After confirmation, add a concise entry to docs/agent-playbooks/known-surprises.md so future agents avoid the same issue. Only record items that are repo-specific, likely to recur, and have a concrete mitigation.

Project Overview

5chan is a serverless, adminless, decentralized 4chan alternative built on the Bitsocial protocol.

Instruction Priority

• MUST rules are mandatory.
• SHOULD rules are strong defaults unless task context requires a different choice.
• If guidance conflicts, prefer: user request > MUST > SHOULD > playbooks.

Task Router (Read First)

Situation	Required action
React UI logic changed (src/components, src/views, src/hooks, UI stores)	Follow React architecture rules below and run yarn doctor
package.json changed	Run yarn install to keep yarn.lock in sync
Translation key/value changed	Use docs/agent-playbooks/translations.md
Bug report in a specific file/line	Start with git history scan from docs/agent-playbooks/bug-investigation.md before editing
CHANGELOG.md or package version changed	Run yarn blotter:check; if needed add a concise release one-liner
UI/visual behavior changed	Verify in browser with playwright-cli; test desktop and mobile viewport
GitHub operation needed	Use gh CLI, not GitHub MCP
User asks for commit/issue phrasing	Use docs/agent-playbooks/commit-issue-format.md
Surprising/ambiguous repo behavior encountered	Alert developer and, once confirmed, document in docs/agent-playbooks/known-surprises.md

Stack

• React 19 + TypeScript
• Zustand for shared state
• React Router v6
• Vite
• plebbit-react-hooks
• i18next
• yarn
• oxlint
• oxfmt
• tsgo

Project Structure

src/
├── components/   # Reusable UI components
├── views/        # Page-level route views
├── hooks/        # Custom hooks
├── stores/       # Zustand stores
├── lib/          # Utilities/helpers
└── data/         # Static data

Core MUST Rules

Package and Dependency Rules

• Use yarn, never npm.
• Pin exact dependency versions (package@x.y.z), never ^ or ~.
• Keep lockfile synchronized when dependency manifests change.

React Architecture Rules

• Do not use useState for shared/global state. Use Zustand stores in src/stores/.
• Do not use useEffect for data fetching. Use plebbit-react-hooks.
• Do not sync derived state with effects. Compute during render.
• Avoid copy-paste logic across components. Extract custom hooks in src/hooks/.
• Avoid boolean flag soup for complex flows; model state clearly in Zustand.
• Use React Router for navigation; no manual history manipulation.

Code Organization Rules

• Keep components focused; split large components.
• Follow DRY: shared UI in src/components/, shared logic in src/hooks/.
• Add comments for complex/non-obvious code; skip obvious comments.

Bug Investigation Rules

• For bug reports tied to a specific file/line, check relevant git history before any fix.
• Minimum sequence: git log --oneline or git blame first, then scoped git show for relevant commits.
• Full workflow: docs/agent-playbooks/bug-investigation.md.

Verification Rules

• Never mark work complete without verification.
• After code changes, run: yarn build, yarn lint, yarn type-check.
• After React UI logic changes, run: yarn doctor.
• Treat React Doctor output as actionable guidance; prioritize error then warning.
• For UI/visual changes, verify with playwright-cli on desktop and mobile viewport.
• If verification fails, fix and re-run until passing.

Tooling Constraints

• Use gh CLI for GitHub work (issues, PRs, actions, dependabot, projects, search).
• Do not use GitHub MCP.
• Do not use browser MCP servers (cursor-ide-browser, playwright-mcp, chrome MCP, etc.).
• Use playwright-cli for browser automation.
• If many MCP tools are present in context, warn user and suggest disabling unused MCPs.

Project Maintenance Rules

• If CHANGELOG.md or package version changes, run yarn blotter:check.
• If blotter check fails for missing release coverage, add a concise one-line release entry.
• Ignore manual entries for release coverage logic.
• Use yarn blotter / yarn blotter:manual for manual dev messages.

Security and Boundaries

• Never commit secrets or API keys.
• Never push to a remote unless the user explicitly asks.
• Test responsive behavior on mobile viewport.

Core SHOULD Rules

• Keep context lean: delegate heavy/verbose tasks to subprocesses when available.
• For complex work, parallelize independent checks.
• When proposing or implementing meaningful code changes, include both:
  • a Conventional Commit title suggestion
  • a short GitHub issue suggestion Use the format playbook: docs/agent-playbooks/commit-issue-format.md.
• When stuck on a bug, search the web for recent fixes/workarounds.
• After user corrections, identify root cause and apply the lesson in subsequent steps.

Local Development URLs

This project uses Portless for local dev. The dev server is available at http://5chan.localhost:1355 instead of a random port. Other Bitsocial projects use the same proxy (seedit, mintpass, bitsocial at .localhost:1355), so they can all run simultaneously without port conflicts.

To bypass Portless: PORTLESS=0 yarn start

Common Commands

yarn install
yarn start                # http://5chan.localhost:1355
yarn build
yarn test
yarn prettier
yarn electron
yarn doctor
yarn doctor:score
yarn doctor:verbose

Playbooks (Load On Demand)

Use these only when relevant to the active task:

• Hooks setup and scripts: docs/agent-playbooks/hooks-setup.md
• Translations workflow: docs/agent-playbooks/translations.md
• Commit/issue output format: docs/agent-playbooks/commit-issue-format.md
• Skills/tools setup and MCP rationale: docs/agent-playbooks/skills-and-tools.md
• Bug investigation workflow: docs/agent-playbooks/bug-investigation.md
• Known surprises log: docs/agent-playbooks/known-surprises.md