Greptile Summary

This PR introduces a /alias and /unalias command system to the TUI, allowing users to save named prompt shortcuts that persist across sessions in a JSON file under the OpenClaw state directory. It also adds a missing vi.mock for dispatchReplyFromConfigWithSettledDispatcher to fix flaky CI runs in the Matrix extension tests.

Key changes:

• src/tui/tui-aliases.ts (new): Alias store with shell-like tokeniser (parseTuiAliasArgs supporting single/double quotes and backslash escapes), name normaliser, JSON file I/O under <state-dir>/tui/aliases.json, and a round-trip sanitiser (normalizeAliasRecord).
• src/tui/tui-command-handlers.ts: Handles /alias [name] [prompt] (list, run, or save) and /unalias <name> (remove). The persistAliases helper uses a correct write-then-update-state pattern so in-memory state is never mutated if the disk write fails.
• src/tui/tui.ts: Loads aliases at startup with a graceful load-failure warning; wires refreshAutocomplete → updateAutocompleteProvider so autocomplete reflects alias changes immediately.
• src/tui/commands.ts: Adds /alias and /unalias to getSlashCommands with prefix-filtered completions and to helpText.
• src/tui/tui-types.ts: Extends TuiStateAccess with tuiAliases: Record<string, string>.
• Test coverage is thorough, including rollback-on-disk-failure and autocomplete refresh assertions.

One minor UX concern: listAliases (invoked by /alias with no arguments) displays each alias as /${name} -> ${prompt}, which may lead users to believe they can invoke it by typing /<name> directly. In practice, unknown slash commands fall through to sendMessage and are sent verbatim to the agent; the correct invocation path is /alias <name>.

Confidence Score: 4/5

• Safe to merge; the alias feature is well-tested and the error-handling pattern is correct. The one flag is a minor display UX issue, not a logic bug.
• The implementation is clean and follows existing patterns in the codebase. The write-before-update state pattern in persistAliases is correct and tested. All new edge cases (empty prompts, reserved names, disk failures, parse errors) have corresponding tests. The only issue is a cosmetic display inconsistency in listAliases where the format /${name} may mislead users about the actual invocation syntax.
• src/tui/tui-command-handlers.ts — the listAliases display format.

This is a comment left during a code review.
Path: src/tui/tui-command-handlers.ts
Line: 264-266

Comment:
**Misleading alias display format**

`listAliases` displays each alias as `/${name} -> ${prompt}`, which implies users can invoke it via `/<name>` (e.g. `/review`). However, the `default` branch of `handleCommand` forwards unknown slash commands straight to `sendMessage`, so typing `/review` would literally send the text `/review` to the agent rather than running the alias. The correct invocation is always `/alias <name>`.

Consider formatting the output so it matches the expected invocation, for example:

```suggestion
    chatLog.addSystem(`/alias ${name} (-> ${prompt})`);
```

Or at a minimum, add a header line clarifying how to invoke them (e.g. `"run with: /alias <name>"`).

How can I resolve this? If you propose a fix, please make it concise.

_{Last reviewed commit: "Handle empty quoted ..."}

Please don't make PRs for test failures on main.

The team is aware of those and will handle them directly on the codebase, not only fixing the tests but also investigating what the root cause is. Having to sift through test-fix-PRs (including some that have been out of date for weeks...) on top of that doesn't help. There are already way too many PRs for humans to manage; please don't make the flood worse.

Thank you.