Skip to content

donaldfilimon/abi / .agents/skills/abi-superpower-tui

abi-superpower-tui

Verified

TUI/dashboard superpower. Launch interactive agent REPL, diagnostics dashboard panes, and slash commands via real abi CLI paths. Use when working on abi tui / dashboard / agent tui. There is no /abi-superpower-tui binary.

FileKindSize
SKILL.mdskill md3.4 KB
---
name: abi-superpower-tui
description: TUI/dashboard superpower. Launch interactive agent REPL, diagnostics dashboard panes, and slash commands via real abi CLI paths. Use when working on abi tui / dashboard / agent tui. There is no /abi-superpower-tui binary.
superpower:
  command: "execute"
  parameters:
    - name: "action"
      type: "string"
      enum: ["repl", "dashboard", "pane", "once", "json"]
      description: "TUI action"
    - name: "pane"
      type: "string"
      description: "Dashboard pane: system, plugins, storage, wdbx, scheduler, memory"
    - name: "compact"
      type: "boolean"
      description: "Render only selected pane"
---

# ABI Superpower: TUI

Exposes the interactive TUI and dashboard as a superpower. There is **no**
`/abi-superpower-tui` binary or slash command — map actions to the real CLI
paths below (or the `run-tui` / `dashboard-smoke` drivers).

## Actions

### repl
Launch the agent REPL (slash commands live here):
```bash
./target/debug/abi agent tui
```

### dashboard
Show the diagnostics dashboard (`abi tui` is an alias of `abi dashboard`):
```bash
./target/debug/abi dashboard --pane system
./target/debug/abi dashboard --pane plugins --once --json
./target/debug/abi tui --compact --pane scheduler
```

Interactive pty smoke (tmux driver):
```bash
.agents/skills/run-tui/tui.sh              # drives `abi dashboard`
.agents/skills/run-tui/tui.sh tui          # drives `abi tui`
```

Headless one-shot smoke:
```bash
.agents/skills/dashboard-smoke/dashboard.sh
```

### pane
Select the initial diagnostics pane with `--pane` (system, plugins, storage/wdbx,
scheduler, memory, or 1–5). In the interactive refresh loop, switch panes with
hotkeys / Tab — there is no separate `pane --focus` CLI.

```bash
./target/debug/abi dashboard --list-panes
./target/debug/abi dashboard --pane memory
```

## Slash Commands (in `abi agent tui` REPL)

- `/help` - Show the REPL command surface
- `/model <id>` - Select a known model or a bounded free-form model ID
- `/profile` - Report the adaptive profile/router state
- `/sea [on|off|status|toggle]` - Change session-local SEA preference only
- `/status` - Report session, model, SEA, and live/store boundaries
- `/context` - Report the bounded local context summary
- `/history` - Print in-session prompt history
- `/reset` - Clear in-session history and turn count
- `/features` - Print locally available feature disclosures
- `/clear` - Clear the terminal display
- `/quit` - Exit (`/exit` and `/q` are aliases)

The REPL does not currently implement `/open`, `/diff`, `/commit`, `/learn`,
`/save`, or `/load`. Evidence-augmented completion is the separate
`abi complete --learn` path.

## Implementation

Maps to:
- `crates/abi-cli/src/repl.rs` - REPL commands, session state, and line editor
- `crates/abi-cli/src/dashboard/` - Split-pane dashboard (`abi tui` / `abi dashboard`)
- `crates/abi-cli/src/terminal.rs` - Raw-mode guard and incremental key decoder
- `.agents/skills/run-tui/tui.sh` - Interactive pty driver
- `.agents/skills/dashboard-smoke/dashboard.sh` - Headless one-shot smoke

## Runtime Boundary

The Rust `abi-cli` build always includes these surfaces. The dashboard refresh
loop and raw editor require a Unix terminal; `--once`, `--json`, and redirected
REPL input remain deterministic non-interactive paths. Dashboard panes support
bounded keyboard navigation and primary-button SGR mouse selection; capture is
disabled by a drop guard before terminal modes are restored.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 323kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 323kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 323kUpdated Claude CodeCodexVerifiedInstall