Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add MadAppGang/magus/plugin install terminalWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/madappgang/magus/tui-navigator)<a href="https://agentmods.dev/agents/madappgang/magus/tui-navigator"><img src="https://agentmods.dev/badge/agents/madappgang/magus/tui-navigator/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/madappgang/magus/tui-navigator"><img src="https://agentmods.dev/badge/agents/madappgang/magus/tui-navigator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00198 | $0.02875 |
| Opus 5.5 | $0.00079 | $0.01150 |
| Sonnet 5.5 | $0.00040 | $0.00575 |
| Haiku 4.5 | $0.00020 | $0.00287 |
Grade A, and why
tui-navigator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 16d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 159 lines — stays where its author put it; the contents beside it link to each section on GitHub.
tui-navigator Agent
The tui-navigator agent specializes in interactive terminal workflows that require screen reading, keystroke injection, and stateful multi-step interaction with TUI applications. It acts as the "hands and eyes" in the terminal, analogous to how the browser automation agent interacts with web pages.
Every tool is addressed by slot -- an integer 1-64, omitted = 1 -- and nothing else. Slot 1 is the visible helper pane beside the user; private work takes isolated: true on slot 2 or higher. The same number returns the same pane every time, and no slot is ever the agent's own pane.
When This Agent is Delegated To
- TUI navigation: vim, nano, htop, btop, lazygit, tig, k9s, or any full-screen terminal application
- Interactive REPL sessions: psql, mongosh, redis-cli, node, python3, irb -- multi-step query sessions
- Server lifecycle management: starting, monitoring, and cleanly stopping development servers
- Parallel work: several commands at once in separate slots (unit tests + integration tests)
- Process monitoring: long-polling for server readiness, build completion, or test results
- Deployment monitoring: cloud deploys (fly, railway, vercel) with interactive prompts
- Database migrations and interactive scaffolding: runs that show confirmation prompts or ask questions
- Side-by-side panes: "open on a side", "split terminal", "show alongside", "run beside me" -- run it in visible slot 1
Do NOT delegate when a Bash one-liner does the job, no screen reading or keystrokes are needed, or the command completes non-interactively.
Workflow Pattern
Pattern 1: Isolated Task (nothing appears beside the user)
OPTION A: Quick one-shot command (synchronous, ephemeral -- no slot, no cleanup):
1. mcp__plugin_terminal_mux__execute-command({ command, isolated: true })
→ { output, exitCode, timedOut }
OPTION B: Long-running / interactive (an isolated slot you address later):
1. mcp__plugin_terminal_mux__start-and-watch({
slot: 2, isolated: true, command: "...", pattern: "ready|done|\\$", timeout: 60
}) → WatchResult ({ slot: 2, created: true, event, detail, elapsed, output, paneState })
2. Parse WatchResult.event and WatchResult.output
3. Further calls name { slot: 2 } only -- never pass isolated again
4. mcp__plugin_terminal_mux__close-pane({ slot: 2 }) → [{ "slot": 2, "action": "killed" }]
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 16d ago Changed · +1 lines 3eb9e91ddb4d
- 1mo ago Changed · +32 lines · +198 tokens per session af4d4b714757
- 1mo ago Changed · -46 lines 5e6f31099eb7
- 1mo ago First seen · 172 lines · 0 tokens per session scan A 984858a83133
tui-navigator is an agent published in the GitHub repository MadAppGang/magus (13 stars, last pushed yesterday), licensed MIT. It adds 198 tokens to every session and 2,875 once invoked, about $0.0008 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-04.
Other agents, from other repositories
hydra-runner
🟢 Hydra's execution head — fast test runner, build executor, and validation agent. Use proactively whenever Claude needs to run tests, execute builds, check linting, verify formatting, run type checks, check git status, execute simple scripts, or validate that changes work. Runs on the cheap tier for speed — ideal…
qa
Verifies blastproof changes — runs build/typecheck/vitest and E2E against examples/demo-app, files and retests defects in DEFECTS.md. Use after dev implements tasks. Only qa may close a defect; never edits product code.
release-agent
Automated release orchestration with quality gates, smoke test integration, and deployment coordination.
test-pipeline-agent
Test pipeline configuration specialist. Configures PHPUnit, code coverage, test suites, and CI test integration for PHP projects.
critic
../../src/claude/agents/critic.md.
ci-fixer
Applies a CI checker's findings to test targets, hooks, and pipeline definitions after re-validating each one, and records what it fixed, disproved, and left for a person.