QM is a shared work environment for coding agents that gives each employee and conversation an isolated workspace while allowing collaboration in Slack and on the web. Startups use it to run agents for individuals, teams, and company operations, with selectable harnesses and models. The catalogue entries extend QM with skills, instructions, and settings.
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 yc-software/qm --skill cloud-cligit clone --depth 1 https://github.com/yc-software/qmWrote 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/yc-software/qm/cloud-cli)<a href="https://agentmods.dev/skills/yc-software/qm/cloud-cli"><img src="https://agentmods.dev/badge/skills/yc-software/qm/cloud-cli/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/yc-software/qm/cloud-cli"><img src="https://agentmods.dev/badge/skills/yc-software/qm/cloud-cli.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 4 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 Privilege Escalation · line 77 Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.Fix: Remove references to credential paths. Use environment variables or secrets managers. For docs, use placeholder paths (e.g., /path/to/config). Never load .env or token files in production code paths.
- high Privilege Escalation · line 87 Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.Fix: Remove references to credential paths. Use environment variables or secrets managers. For docs, use placeholder paths (e.g., /path/to/config). Never load .env or token files in production code paths.
- medium Agent Snooping · line 29 Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
- medium Agent Snooping · line 64 Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
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 | $0.00059 | $0.01195 |
| Opus 5.5 | $0.00024 | $0.00478 |
| Sonnet 5.5 | $0.00012 | $0.00239 |
| Haiku 4.5 | $0.00006 | $0.00120 |
Grade A, and why
cloud-cli 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 3d 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 — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Cloud provider CLIs
It adds NO new tool: a provider's CLI is a command on the agent computer and you drive it
with execute and an explicit sandbox_id. Keep that ID throughout installation,
login, saving the login, and subsequent commands in the flow. Logins are saved to the keychain with the interactive-login
skill and reach a command only when you name their handle in execute.credentials.
First: is the CLI even here?
Do not assume the binary is on PATH. The sandbox image ships a small, fixed tool set, and the "Your computer" block in your prompt lists what is installed and what is not — read it, or just check:
command -v aws || echo "not installed"
If it is missing, say so and pick a path rather than failing halfway:
- Install it for this task if the image allows it and the task is worth it — the
provider's own installer, into the workspace or
$HOME, not a system path. Say that you installed it; it lasts as long as the machine's disk. - Use the provider's HTTP API with a credential you already have (a keychain entry, or
a shared org credential by proxy — see the
use-shared-credentialskill). - Ask the operator to add the CLI to the sandbox image if this will recur. That is the durable fix; a per-turn install is not.
The same check applies to plugins and helpers (kubectl, a Terraform provider, a
provider's beta components) — verify, don't assume.
Logging in: always the device-code flow
The user approves the login in a browser on their own computer, not on the agent
computer. Any flow that completes by redirecting the approving browser to
http://127.0.0.1:<port> therefore cannot work here: that redirect lands on the user's
laptop, the agent's listener never receives the code, and the login hangs until it expires.
Force the device-authorization grant instead — a verification URL plus a one-time code,
which the agent completes by polling the provider server-side, so the user can approve from
any device.
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.
- 3d ago Changed · +1 lines 31b885c27e40
- 7d ago Changed · -7 lines · -4 tokens per session 535d68915669
- 19d ago Changed · +9 lines ae0a2c2e6bcd
- 27d ago Changed 41baf3bda202
- 1mo ago Changed 33eb5d95ef08
- 1mo ago First seen · 89 lines · 63 tokens per session scan A dd50f1129b42
cloud-cli is a skill published in the GitHub repository yc-software/qm (15,434 stars, last pushed today), licensed MIT. It adds 59 tokens to every session and 1,195 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
cloud-image-bake
Bake, select, prove, and safely retire a Cloud Worker image with crabbox and config one-liners.
cloudflare-skill
Work with Cloudflare via the official cf CLI — find the right cf command, list zones, list/create/edit/delete DNS records, query the GraphQL Analytics API (zone traffic, Web Analytics/RUM), and run any other cf command (Workers, R2, KV, D1, Pages, cache). Output is parsed JSON.
gcloud-skill
Work with Google Cloud via the official gcloud CLI — check auth/config, list and switch projects, manage Compute Engine instances, deploy and inspect Cloud Run services, work with Cloud Storage, and run any other gcloud command. Output is parsed JSON.
vercel-skill
Deploy sites and apps to Vercel, inspect deployments, stream logs, and manage projects/env/domains via the Vercel CLI. Deploy a directory and get back the live deployment URL; everything else the CLI supports is available through the custom command passthrough.
aws-skill
Work with Amazon Web Services through the official boto3 SDK — check the identity behind the key, list, start and stop EC2 instances, list, upload, download and delete S3 objects, and call any other AWS API operation. Output is AWS's own JSON response.
web-design
A tool for designing and publishing single-page HTML websites, such as reports, landing pages, and data displays. It uses common web libraries for layouts, charts, icons, and responsive pages, then can deploy the result to Cloudflare Pages.