devicecloud-mcp AGENTS.md

Instructions for using devicecloud-mcp to investigate failed DeviceCloud mobile test runs. DeviceCloud runs Maestro flows, which are automated mobile-app tests, and the service provides failure details, screenshots, logs, and flow history to an AI coding agent.

In plain words
What is it for?
Use it when CI reports a failed Maestro run on DeviceCloud and you need the failure reason, screenshots, logs, or execution history to diagnose it.
Why use it?
It removes the need to open the DeviceCloud dashboard to gather debugging evidence. The agent can inspect the failed run in the editor, identify the likely cause, and then help fix the flow or app code.

Instructions file for CodexOpenCode

Install

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.

agentmods
npx agentmods add instructions/rubenglez/devicecloud-mcp/agents-md
Clone the repo
git clone --depth 1 https://github.com/RubenGlez/devicecloud-mcp

Made for: Codex, OpenCode.

Per session 1,002 This file is loaded in full into every session.
When invoked 1,002 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.01002 $0.01002
Opus 5 $0.00501 $0.00501
Sonnet 5 $0.00200 $0.00200
Haiku 4.5 $0.00100 $0.00100

Measured 2d ago against content hash 6e578b925ca9, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

devicecloud-mcp AGENTS.md 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 2d 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.

AGENTS.md · 62 lines

How it starts

The opening of the file, as written. The whole thing — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.

devicecloud-mcp

Triage failing DeviceCloud runs so your agent can fix them, without opening the dashboard.

The dashboard is no longer in your debug loop. When CI goes red, your agent pulls the failure straight into your editor (fail reasons, failure screenshots, logs, and flow history), finds the root cause, and you commit the fix. CI re-runs. You never open a browser tab.

Who it's for

Mobile teams running Maestro flows on DeviceCloud through CI, who debug failures with an AI coding agent (Claude Code, Cursor) instead of the web dashboard.

The problem

When a CI run fails, the only way to understand why lives in the DeviceCloud dashboard: open it, click into the run, read the fail reason, study the failure screenshot, cross-reference the logs, then switch back to your editor to fix the flow. The evidence you need to fix the test is stuck in a browser, away from the code and the agent that could act on it.

The promise

The MCP delivers the triage; your agent does the fix. One invocation turns a red run into root cause plus the exact artifacts to act on, your agent edits the flow or app code, and you commit. The dashboard stays closed.

Why it's different

  • vs. the DeviceCloud dashboard: the dashboard shows a human why a run failed; this lets an agent read the same evidence and act on it: edit the flow, not just look at the screenshot.
  • vs. the official DeviceCloud MCP: that ships a handful of thin endpoint wrappers built around running tests. This is triage-first: opinionated tools that hand the agent root cause and the right artifacts, not raw JSON to stitch together. Running isn't its job; CI owns that.
  • vs. the Maestro MCP: that drives local devices to author and run tests. This is cloud-only and post-run: it works on the runs CI already executed.

Scope boundaries (deliberate)

  • Doesn't trigger or re-run tests; CI plus git commit owns that.
  • No dcd CLI dependency; pure REST API, npx-and-go, read-only and safe (an agent can't burn credits).
  • Not a local-device or test-authoring tool.
  • Cancelling a run is dashboard-only and outside the triage loop this is built for.

Read the full file on GitHub · 62 lines

Changes

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.

  1. 2d ago First seen · 62 lines · 1,002 tokens per session scan A 6e578b925ca9

Subscribe to this mod's changes

devicecloud-mcp AGENTS.md is an instructions file published in the GitHub repository RubenGlez/devicecloud-mcp (0 stars, last pushed 1mo ago), licensed MIT. It adds 1,002 tokens to every session, about $0.0050 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 2026-08-31.