matterbridge-eve-energy

Imported from Luligu/matterbridge-eve-energy (AGENTS.md). Install upstream with npx skills add Luligu/matterbridge-eve-energy. Copyright stays with the author.

Matterbridge Agents Instructions (v.1.0.7)

Style And Formatting

• Follow STYLEGUIDE.md for code style, naming, JSDoc, validation, logging, and formatting expectations.
• JSDoc requirements are enforced by the configured linter. Treat missing or incomplete JSDoc on required APIs as a real lint issue, not optional documentation.
• Import and export ordering are enforced by the formatter. Preserve the existing grouped and sorted order unless a change requires updating it.
• Formatting is enforced by oxfmt. Follow the existing formatting and do not fight the formatter.

Scope And Safety

• Keep changes minimal and scoped to the request. Avoid unrelated refactors or broad cleanup.
• Do not modify production code only to make a test pass. If a failing test points to a likely source issue, explain the issue and change behavior only when required by the task.
• Preserve cross-platform behavior. Changes must work on Windows, macOS, and Linux, especially for paths, shell commands, environment variables, and networking behavior.
• Maintain compatibility with the supported Node.js versions in this repository: 20.19, 22.13, 24 and 26.

Project Architecture

• This repository uses TypeScript and ESM. Follow existing project patterns for imports, exports, build configuration, and test setup.

Testing And Validation

• HARD RULE: never invoke tsc, vitest, bun test, oxlint, or oxfmt directly (via npx, node node_modules/..., or any other ad hoc form). Always use the matching entry in package.json scripts or tasks.json — e.g. npm run typecheck (not npx tsc), npm run lint (not npx oxlint; npm run lint:fix rewrites files, so run it only when the user asks), npm run format/npm run format:check (not npx oxfmt), npm run test/npm run test:coverage or a specific Test: <Area> task (not npx vitest/node node_modules/vitest/vitest.mjs typed out by hand), npm run test:bun (not bun test). If a touched area has no matching script or task, say so and ask before improvising a raw invocation — do not silently fall back to npx.
• Keep tests deterministic and simple. Prefer small data sets and straightforward setup.
• Some tests are intentionally multi-step flows. State may persist across successive steps within a single test flow, but each test unit must remain isolated from other tests.
• For validation, run the relevant full test file or the matching suite/task for the touched area rather than assuming arbitrary isolated single-test execution is reliable.

Documentation

• When behavior changes, update the relevant tests and documentation.

Additional Agent Guidance

For task-specific guidance, read relevant files in .agents/rules:

• .agents/rules/testing.instructions.md for testing and validation expectations;
• .agents/rules/matterbridge.instructions.md for guidance on creating and registering Matterbridge endpoints in a plugin;
• .agents/rules/plugin-frontend.instructions.md for guidance on plugin frontend SPAs and custom REST APIs;
• .agents/rules/chip-tests.instructions.md for guidance on the CHIP conformance test harness.

The following workflows are available as skills in .agents/skills, discovered automatically and invocable with $<name> (Codex) or /<name> (Claude Code, Copilot, Gemini / Antigravity):

• $verify-agent-context for verifying which coding agent is running and that it loaded the shared instructions, rules and skills from AGENTS.md and .agents/.