elizaos--eliza/plugins/plugin-native-messages/AGENTS.md

@elizaos/capacitor-messages

Capacitor plugin that gives an Eliza agent on Android the ability to send and read native SMS/MMS messages via the Android Telephony API.

Purpose / role

This is a Capacitor plugin — not an elizaOS Plugin object. It bridges the Android SmsManager and content://sms provider into a typed JavaScript API. It is consumed by any Capacitor-based Eliza app running on Android; the web fallback throws on sendSms and returns an empty list from listMessages. It is opt-in: register it in your Capacitor Android project, request the required runtime permissions, and import Messages from @elizaos/capacitor-messages.

Plugin surface

This plugin does not register elizaOS actions, providers, evaluators, or services. It exposes a Capacitor plugin interface named ElizaMessages with two methods:

Method	Description
Messages.sendSms({ address, body })	Sends an SMS (multipart if needed); waits for radio confirmation; persists to Android sent folder. Returns { messageId, messageUri }.
Messages.listMessages({ limit?, threadId? })	Reads up to limit messages (default 100, max 500) from the system SMS store, optionally filtered by threadId. Returns { messages: SmsMessageSummary[] }.

Layout

plugins/plugin-native-messages/
  src/
    index.ts          Entry point — calls registerPlugin("ElizaMessages", { web: loadWeb })
    definitions.ts    TypeScript interfaces: MessagesPlugin, SendSmsOptions, SendSmsResult,
                      ListMessagesOptions, SmsMessageSummary
    web.ts            Web fallback — sendSms throws; listMessages returns []
    web.test.ts       Vitest unit tests for the web fallback
  android/
    src/main/
      AndroidManifest.xml               READ_SMS / SEND_SMS / RECEIVE_SMS / RECEIVE_MMS / RECEIVE_WAP_PUSH permission declarations
      java/ai/eliza/plugins/messages/
        MessagesPlugin.kt               Capacitor @CapacitorPlugin("ElizaMessages"); implements
                                        sendSms (SmsManager + BroadcastReceiver delivery receipt)
                                        and listMessages (ContentResolver query on content://sms)
  rollup.config.mjs   Bundles dist/esm → dist/plugin.js (IIFE) and dist/plugin.cjs.js
  tsconfig.json
  package.json

Commands

Scripts are defined in package.json; run them from the repo root with bun run --cwd:

bun run --cwd plugins/plugin-native-messages clean           # remove build output
bun run --cwd plugins/plugin-native-messages build           # build package artifacts
bun run --cwd plugins/plugin-native-messages typecheck       # TypeScript typecheck
bun run --cwd plugins/plugin-native-messages lint            # mutating Biome check
bun run --cwd plugins/plugin-native-messages lint:check      # read-only Biome check
bun run --cwd plugins/plugin-native-messages format          # write formatting
bun run --cwd plugins/plugin-native-messages format:check    # read-only formatting check
bun run --cwd plugins/plugin-native-messages test            # run package tests
bun run --cwd plugins/plugin-native-messages prepublishOnly  # publish-time build hook
bun run --cwd plugins/plugin-native-messages build:unlocked  # bun run clean && tsc && bunx rollup -c rollup.config.mjs

Config / env vars

No environment variables or elizaOS config keys. The plugin reads no .env values. All behaviour is determined at call time:

• SEND_SMS Android runtime permission — required for sendSms.
• READ_SMS Android runtime permission — required for listMessages.

Both permissions are declared in android/src/main/AndroidManifest.xml. The host app must request them at runtime before calling either method.

How to extend

Add a new Capacitor method (e.g., deleteSms):

1. Add the method signature to MessagesPlugin in src/definitions.ts.
2. Add a web fallback in src/web.ts that throws "deleteSms is only available on Android.".
3. Implement @PluginMethod fun deleteSms(call: PluginCall) in android/src/main/java/ai/eliza/plugins/messages/MessagesPlugin.kt using the ContentResolver.
4. If the new method needs an Android permission, declare it in AndroidManifest.xml and check it with hasPermission(Manifest.permission.*) before proceeding.
5. Run bun run --cwd plugins/plugin-native-messages build to regenerate dist/.

Conventions / gotchas

• Android only. The web fallback exists solely to satisfy Capacitor's plugin registration contract. Do not add real web logic here.
• Instrumented test (issue #9967). The content://sms query lives in MessagesReader; an on-device read test (android/src/androidTest/.../MessagesReaderInstrumentedTest.kt, GrantPermissionRule) reads back a marker SMS. It is emulator-orchestrated (adb -s <emulator> emu sms send <number> "…Eliza-9967-SMS-roundtrip…" then connectedDebugAndroidTest/am instrument) and Assume-skips when the marker is absent, so it never reads a real device's private inbox. listMessages delegates to the reader (JS shape unchanged).
• Multipart SMS. sendSms uses SmsManager.divideMessage and tracks one BroadcastReceiver delivery intent per part; the call resolves only after all parts confirm. Do not assume a single sendTextMessage call for long messages.
• Delivery receipt vs. sent receipt. The BroadcastReceiver listens for SENT status only. Delivery receipts (DELIVERED) are not tracked.
• Limit clamp. listMessages rejects if limit is outside [1, 500]. The Android SMS provider can be large; do not request unbounded results.
• Plugin name. The Capacitor plugin name is "ElizaMessages" (set in both index.ts and the Kotlin @CapacitorPlugin annotation). The npm package name is @elizaos/capacitor-messages. The directory is plugin-native-messages. All three differ — keep them in sync if renaming.
• Build output. tsc emits to dist/esm/; rollup then bundles dist/esm/index.js into dist/plugin.js (IIFE) and dist/plugin.cjs.js. The exports field in package.json uses dist/esm/index.js for ESM consumers and dist/plugin.cjs.js for CJS.
• Peer dep. @capacitor/core ^8.3.1 is a peer dependency; the host Capacitor app owns the exact version.

⛔ NON-NEGOTIABLE — evidence, trajectories & real end-to-end tests

The binding, repo-wide standard is AGENTS.md. Read it. Nothing in this package is done until it is proven done — a reviewer must confirm it works without reading the code, from the artifacts you attach. This applies to every feature, fix, refactor, and chore here. "Tests pass" is not proof; "CI is green" is not proof.

• Record AND read model trajectories. Capture the actual inputs and outputs of the model from a live LLM — not the deterministic proxy, not a mock: the prompt, the providers/context, the raw model output, every tool/action call, and the result. Then open the trajectory and review it by hand. A captured-but-unread trajectory is not evidence (packages/scenario-runner/bin/eliza-scenarios run <scenario> --report <out>).
• Real, full-featured E2E — no larp. Every feature ships detailed end-to-end tests that drive the real path end to end. Not the happy "front door" only: cover error paths, edge/empty/invalid input, concurrency, roles/permissions, and adversarial input. A test that asserts against a mock/stub/fixture standing in for the thing under test does not count. If the real model/device/chain/connector/account is hard to reach, make it reachable — that is the work, not an excuse to mock. If the existing tests here are shallow or mocked, fixing them is part of your change.
• Screenshots + logs at every phase, plus a complete walkthrough video/run-through of the entire feature or view, start to finish (bun run test:e2e:record).
• Manually review every artifact the change touches — never just the green check: client logs (console + network), server logs ([ClassName] …), the model trajectories in and out, before/after full-page screenshots, and the domain artifacts listed below for this package.
• No residuals. No shortcuts. The goal is not "done" — it is everything done. Clear every blocker by the hard path: build the real architecture, stand up the real model/device/service, actually test it. Never leave a TODO, a stub, a stepping-stone, or a "follow-up." When unsure, research thoroughly, weigh the options, and ship the best, highest-effort, production-ready version. Keep going until every possibility is exhausted.

Artifacts → attached inline in the PR (MP4 video, JPG screenshots, logs in <details>); attach each evidence type or explicitly mark it N/A with a reason — never leave it blank. If develop moved and changed behavior, re-capture evidence; stale proof is worse than none.

Capture & manually review for this package — native / on-device bridge:

• The capability run on a real device or simulator — not desktop Chromium against a mocked bridge (see #9967/#9580): device logs + the captured output (photo, OCR text, detection boxes, transcript, sensor reading).
• Parity vs the reference implementation where one exists (e.g. the Python/Ultralytics reference), with the numeric tolerances actually met.
• Permission-denied, no-hardware, and background/foreground lifecycle paths.
• A short recording of the on-device run; confirm the build under test is yours (versionName / a known on-screen change), not a stale install.