Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add kobiton/automate --skill drive-automation-sessiongit clone --depth 1 https://github.com/kobiton/automateWrote 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/skills/kobiton/automate/drive-automation-session)<a href="https://agentmods.dev/skills/kobiton/automate/drive-automation-session"><img src="https://agentmods.dev/badge/skills/kobiton/automate/drive-automation-session/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/skills/kobiton/automate/drive-automation-session"><img src="https://agentmods.dev/badge/skills/kobiton/automate/drive-automation-session.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 3 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Tool Misuse · line 173 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
- high Tool Misuse · line 328 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
- medium Excessive Agency · line 101 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00150 | $0.05992 |
| Opus 5 | $0.00075 | $0.02996 |
| Sonnet 5 | $0.00030 | $0.01198 |
| Haiku 4.5 | $0.00015 | $0.00599 |
Grade A, and why
drive-automation-session 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 11d 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 — 367 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Drive from Intent
Overview
Drive a Kobiton device from a natural-language intent. The caller provides:
- An intent string (e.g. "open Settings, enable Bluetooth, then go back to home").
- A UDID for a device the caller has already reserved.
- An app reference (
kobiton-store:vXXXXX) or a browser name for web sessions.
The skill opens an automation-type Appium session (with appium:newCommandTimeout: 1800 so the session survives pauses while the model asks for guidance), runs an observe-decide-act loop, and returns the session id. The returned session is consumable by the Kobiton MCP tools getSession, getSessionArtifacts, and saveTestCase exactly like any other automation session.
Tool naming. This doc refers to Kobiton MCP tools by their bare names (
getSession,reserveDevice,terminateSession, …). The exact registered name depends on how the host loaded the MCP server (e.g. Claude Code as a plugin exposesmcp__plugin_automate_kobiton__getSession; a repo-local.mcp.jsonorclaude mcp addexposesmcp__kobiton__getSession; other hosts differ). Models fuzzy-match on the bare name, so use it and let the host resolve the prefix.
This skill complements — and does NOT replace — run-interactive-session. They serve different session types:
| Skill | Session type | Best for |
|---|---|---|
run-interactive-session |
CLI session (~/.kobiton/bin/kobiton) |
Human-driven exploration, ad-hoc inspection |
drive-automation-session |
Automation session (direct Appium HTTP) | AI-driven flows, saveable as a test case |
Prerequisites
Runs on any OS, but needs two things from the host: a persistent local filesystem — the observe-decide-act loop passes state between turns through iter-N.*.json files, so that file handoff is the loop — and ~/.kobiton/.credentials, which scripts/appium.js reads directly (see below). A host with a filesystem but no way to run /automate:setup still can't finish; route those users to create-test-run. Node.js 18+; no native binary (scripts/appium.js uses node:https). See the Skill compatibility matrix in CLAUDE.md.
What ships with it
8 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/capabilities.md 4.5 KB
- references/endpoint-reference.md 22 KB
- references/loop-discipline.md 16 KB
- scripts/__fixtures__/webview-mobile-sample.xml 50 KB
- scripts/appium.js 17 KB runs code
- scripts/appium.test.js 28 KB runs code
- scripts/strip-webview-dom.js 2.7 KB runs code
- scripts/strip-webview-dom.test.js 4.9 KB runs code
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.
- 11d ago First seen · 367 lines · 150 tokens per session scan A 9efc4959264a
drive-automation-session is a skill published in the GitHub repository kobiton/automate (12 stars, last pushed 7d ago), licensed MIT. It adds 150 tokens to every session and 5,992 once invoked, about $0.0007 per session on Opus 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 2026-08-30.
Other skills, from other repositories
Appium Mobile Testing
Mobile application testing skill using Appium for iOS and Android, covering device capabilities, selectors, gestures, and cross-platform testing strategies.
playwright-for-android
End-to-end Android debug loop on an emulator. Builds the APK with a Gradle task, boots a named AVD, installs and launches the target package, then drives a single feature via adb-injected UI actions while capturing a unified logcat-and-actions timeline plus a developer-facing test report. Use when the user wants to…
android-ui-journey-testing
XML-specified Android UI journey testing, interactive step execution, assertion verification, and JSON outcome reporting.
solopi-ai
A command-line framework for testing Android apps and devices with SoloPi, including on-device or cloud AI decision models. It manages devices, test cases, recorded interactions, replays, performance history, and evidence.
android-emulator
Verify and debug native, React Native, Expo, or Flutter apps on an Android Emulator with agent-device. Use when an agent needs to launch an app, inspect its live UI, tap, type, scroll, validate a code change, collect failure evidence, or reproduce a workflow on an Android virtual device.
dogfood
Systematically explore and test a mobile app on iOS/Android with agent-device to find bugs, UX issues, and other problems. Use when asked to dogfood, QA, exploratory test, find issues, bug hunt, or test this app on mobile.