# OpenClaw Patch Mapping Matrix

## Purpose

Map the current runtime patch behaviors in `openclawtest` to the repo-side modules that should eventually own them.

This is the bridge between PoC patching and plugin-oriented implementation.

## Matrix

| Current runtime behavior | Current patched file | Intended repo-side owner | Current repo status | Notes |
|---|---|---|---|---|
| Default reply-end button injection | `send-sxDwUGaO.js` | `src/runtime/openclaw-telegram-poc.ts` + `src/telegram/reply-decorator.ts` + `src/runtime/openclaw-telegram-bridge.ts` | strong | Default buttons now exist as Telegram helper and bridge helper; tested repo-side |
| Callback data parsing (`rec:continue`, `rec:stop`) | `bot-Ce301bOE.js` | `src/core/callback-contract.ts` + `src/telegram/callback-handler.ts` | strong | Parsing and normalization logic exist and are tested |
| Resolved button state UI (`✅` / `⏹`) | `bot-Ce301bOE.js` | `src/runtime/openclaw-telegram-poc.ts` + `src/runtime/openclaw-telegram-bridge.ts` + `src/runtime/openclaw-telegram-callback-actions.ts` | strong | Helper chain now owns the resolved-button model |
| Acknowledgement reply text | `bot-Ce301bOE.js` | `src/runtime/openclaw-telegram-poc.ts` + `src/runtime/openclaw-telegram-bridge.ts` + `src/runtime/openclaw-telegram-callback-actions.ts` | strong | Ack text now has explicit repo-side helper ownership and tests |
| State file persistence (`reply-end-controls.json`) | `bot-Ce301bOE.js` | `src/core/store.ts` + `src/adapters/openclaw-state-file.ts` | strong | Store and adapter helpers exist and have tests; runtime patch still writes directly instead of delegating |
| Stop-state inbound behavior injection | `bot-Ce301bOE.js` | `src/runtime/openclaw-telegram-poc.ts` + `src/core/policy.ts` + `src/runtime/openclaw-telegram-bridge.ts` | good | Policy behavior now has repo-side helper coverage; runtime still injects literal text inline |
| Callback debug logging | `bot-Ce301bOE.js` | future adapter/runtime diagnostics module | missing | Still PoC-only runtime instrumentation with no clean repo-side abstraction yet |

## Immediate migration priority

The best first migration target is:

### Priority 1

Move button construction and acknowledgement text to repo-owned helper use.

Why:

- smallest patch surface
- easiest to verify
- repo-side helper and bridge now both exist

### Priority 2

Move state persistence to explicit adapter helper use.

Why:

- store code already exists
- adapter bridge already exists
- removes fragile duplicated inline state-writing logic

### Priority 3

Move stop-state behavior out of large inline injected strings into clearer policy-driven adapter logic.

Why:

- highest behavior importance
- also highest risk if changed too early

## Current assessment

The repo now contains meaningful building blocks for the future pluginized path, including tested core/store/helper/bridge/callback-resolution modules, but the live OpenClaw runtime still contains direct inline patch logic.

That means pluginization has started, but the current implementation is still in a hybrid state.