Install the "cosa-platforms" agent skill from https://github.com/coreos/coreos-assembler/tree/main/.agents/skills/cosa-platforms into .claude/skills/cosa-platforms/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cosa-platforms", then confirm the skill loads.
Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Type this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
skills CLI
$ npx skills add coreos/coreos-assembler --skill cosa-platforms -a codex
Project install goes to .agents/skills/; add -g for ~/.codex/skills/.
Install the "cosa-platforms" agent skill from https://github.com/coreos/coreos-assembler/tree/main/.agents/skills/cosa-platforms into .agents/skills/cosa-platforms/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cosa-platforms", then confirm the skill loads.
Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
skills CLI
$ npx skills add coreos/coreos-assembler --skill cosa-platforms -a cursor
Project install goes to .agents/skills/; add -g for ~/.cursor/skills/.
Install the "cosa-platforms" agent skill from https://github.com/coreos/coreos-assembler/tree/main/.agents/skills/cosa-platforms into .cursor/skills/cosa-platforms/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cosa-platforms", then confirm the skill loads.
Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
skills CLI
$ npx skills add coreos/coreos-assembler --skill cosa-platforms -a gemini-cli
Project install goes to .agents/skills/; add -g for ~/.gemini/skills/.
Install the "cosa-platforms" agent skill from https://github.com/coreos/coreos-assembler/tree/main/.agents/skills/cosa-platforms into .gemini/skills/cosa-platforms/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cosa-platforms", then confirm the skill loads.
Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Installs for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
skills CLI
$ npx skills add coreos/coreos-assembler --skill cosa-platforms -a github-copilot
Project install goes to .agents/skills/; add -g for ~/.copilot/skills/.
Install the "cosa-platforms" agent skill from https://github.com/coreos/coreos-assembler/tree/main/.agents/skills/cosa-platforms into .github/skills/cosa-platforms/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cosa-platforms", then confirm the skill loads.
GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
skills CLI
$ npx skills add coreos/coreos-assembler --skill cosa-platforms -a opencode
OpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
Install the "cosa-platforms" agent skill from https://github.com/coreos/coreos-assembler/tree/main/.agents/skills/cosa-platforms into .opencode/skills/cosa-platforms/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cosa-platforms", then confirm the skill loads.
OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
Skill name
cosa-platforms
GitHub stars
395
Token cost
~1.6k tokens
SKILL.md length
694 words
Files
1
Skills in repo
4
Repo updated
First seen
Licence
Apache-2.0
At a glance
Build disk images, upload to cloud providers, and run tests or spawn instances on non-QEMU platforms (AWS, GCP, Azure, ISO, metal, etc.)
Works in 5 steps: Build the platform image → Upload to the cloud provider → Set up credentials → …
SKILL.md covers Related skills, Key documentation references, Supported platforms and Workflow, plus 1 more section
Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Cosa Platforms is an agent skill from coreos/coreos-assembler. Build disk images, upload to cloud providers, and run tests or spawn instances on non-QEMU platforms (AWS, GCP, Azure, ISO, metal, etc.)
Its SKILL.md is about 1.6k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
It works with Amazon Web Services, Google Cloud and Microsoft Azure. The repository describes itself as: Tooling container to assemble CoreOS-like systems. The licence is Apache-2.0.
Example prompts
“/cosa-platforms”
Workflow steps
5 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit 4ab8b08. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Cosa Platforms loads about 1.6k tokens when it runs. Until then it costs about 38 tokens; SKILL.md has 694 words of instructions outside code blocks.
Always· name and description, kept in context so the agent knows when to use it
~38
When it runs· the whole SKILL.md, loaded when a task matches
~1.6k
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check: warnings
The automated check found patterns that need a careful read before installing.
WarningMentions a credentials file (SSH keys, cloud or package-manager tokens)SKILL.md:41
WarningMentions a credentials file (SSH keys, cloud or package-manager tokens)SKILL.md:129
` | Inject SSH keys from the agent and `~/.ssh/` |
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
Download SKILL.mdSave it as .claude/skills/cosa-platforms/SKILL.md (or your agent's skills folder).
name
cosa-platforms
description
Build disk images, upload to cloud providers, and run tests or spawn instances on non-QEMU platforms (AWS, GCP, Azure, ISO, metal, etc.)
Platform Images and Cloud Testing with cosa
This skill covers working with non-QEMU platforms in coreos-assembler:
building platform-specific disk images, uploading them to cloud providers,
running kola tests on cloud platforms, and spawning interactive cloud
instances.
Related skills
Load the cosa-basics skill for workdir setup, the cosa() bash alias,
and general build workflow.
Load the cosa-kola skill for general kola test mechanics (test
discovery, result inspection, common flags, external tests).
This skill adds the platform-specific details on top of those.
Key documentation references
Do NOT duplicate these docs. Read them at runtime for command details:
File
What it covers
docs/mantle/credentials.md
Credential setup for all platforms (AWS, Azure, GCP, OpenStack, DO, ESX, etc.)
docs/cosa.md
Command reference including osbuild and imageupload-*
docs/cosa/imageupload-aws.md
AWS AMI upload (IAM permissions, credentials)
docs/cosa/osbuild-secex.md
IBM Secure Execution image building
docs/cosa/run.md
cosa run (QEMU-only) and kola spawn for cloud instances
docs/kola.md ("Run tests on cloud platforms")
Cloud testing examples for AWS, GCP, Azure
mantle/cmd/kola/options.go
Source of truth for all platform-specific flags
Supported platforms
Kola supports the following platforms via the -p/--platform flag:
Platform
-p value
Image flag
Default credentials
AWS
aws
--aws-ami
~/.aws/credentials
Azure
azure
--azure-disk-uri
~/.azure/azureCreds.json
DigitalOcean
do
--do-image
~/.config/digitalocean.json
ESX/VMware
esx
--esx-base-vm
~/.config/esx.json
GCP
gcp
--gcp-image
GCP service account JSON key
OpenStack
openstack
--openstack-image
~/.config/openstack.json
QEMU
qemu
--qemu-image
(none)
QEMU ISO
qemu-iso
--qemu-iso
(none)
Read docs/mantle/credentials.md for full credential setup details per
platform, and mantle/cmd/kola/options.go or cosa kola run -h for the
complete flag list.
Workflow
Step 1: Build the platform image
cosa osbuild <platform>
Use cosa osbuild --supported-platforms to list all available platforms.
Multiple platforms can be built in one invocation:
cosa osbuild aws azure gcp
This creates disk images in the appropriate format for each platform (e.g.
VMDK for AWS, VHD for Azure, tar.gz for GCP, raw for metal).
Step 2: Upload to the cloud provider
For cloud platforms that require image upload:
cosa imageupload-<platform>
Available upload targets: aliyun, aws, azure, gcp, powervs.
Read docs/cosa/imageupload-aws.md for AWS-specific IAM and credential
setup. Not all platforms have an imageupload-* command; for those that
don't, use ore directly or upload manually.
Step 3: Set up credentials
Read docs/mantle/credentials.md for per-platform credential configuration.
Each platform has its own credential file format and default location (see
the table above).
Step 4: Run tests on a cloud platform
cosa kola run -p <platform> [platform-specific flags] <test-pattern>
Examples:
# AWS
cosa kola run -p aws --aws-ami <AMI> --aws-region us-east-1 basic
# GCP
cosa kola run -p gcp --gcp-image <image> --gcp-json-key <key> --gcp-project <project> basic
# Azure
cosa kola run -p azure --azure-credentials <file> --azure-disk-uri <URI> basic
Read docs/kola.md ("Run tests on cloud platforms") for more examples.
For the full list of flags per platform, read mantle/cmd/kola/options.go
or run cosa kola run -h.
Show full SKILL.md (303 more words)Show less
Step 5: Spawn instances for interactive debugging
cosa kola spawn -p <platform> [platform-specific flags]
This launches an instance and opens an SSH shell without running tests.
Useful for interactive debugging. Key flags:
Flag
Purpose
--shell / -s
Open an SSH shell (default: true)
--detach / -t
Spawn in the background (implies --shell=false --remove=false --verbose --keys)
--remove=false
Keep the instance running after the shell exits
--keys / -k
Inject SSH keys from the agent and ~/.ssh/
--nodecount / -c
Number of instances to spawn
--userdata / -u
Path to Ignition/Butane config for the instance
--ssh-command / -x
Run a command over SSH instead of opening a shell
Note: --reconnect is only available on QEMU platforms.
Read docs/cosa/run.md ("Running on cloud platforms") for more details.
Important notes
cosa run is QEMU-only. For all other platforms, use
cosa kola run -p <platform> (for tests) or
cosa kola spawn -p <platform> (for interactive instances).
When a cosa build exists in the workdir, kola auto-fills platform image
references from the build metadata (e.g. AMI for AWS, image name for GCP).
This means you may not need to specify --aws-ami etc. if the build
already has that info in meta.json. Check builds/<id>/<arch>/meta.json
to see if the cloud image info is present.
If images already exist in the cloud for a build, you can skip both
cosa osbuild and cosa imageupload-* entirely. Just fetch the build
metadata with a lightweight cosa buildfetch (no --artifact or
--decompress needed):
cosa buildfetch -b <build-id>
Without --artifact, this only downloads a few small metadata files
(meta.json, commitmeta.json, etc.) -- not any disk images. The
meta.json contains the cloud image references (AMI IDs, GCP image
names, etc.) and kola will pick them up automatically when you run
cosa kola run -p <platform>.
Instance types default to reasonable values per platform and architecture
(e.g. m5.large for AWS x86_64, c6g.xlarge for aarch64). Override with
platform-specific type flags (e.g. --aws-type, --gcp-machinetype,
--azure-size).
Cosa Platforms next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
Deep static code review of an mql provider for logic errors, nil-handling bugs, pagination truncation, caching/id collisions, and other defects that silently give users wrong data.
Develop and test changes to coreos-assembler itself -- build mantle Go binaries (kola, ore, plume), iterate on Python/shell scripts, or do full container rebuilds
Build disk images, upload to cloud providers, and run tests or spawn instances on non-QEMU platforms (AWS, GCP, Azure, ISO, metal, etc.). Cosa Platforms is an agent skill from coreos/coreos-assembler.
How do I install Cosa Platforms in Claude Code?
Run `npx skills add coreos/coreos-assembler --skill cosa-platforms -a claude-code`. Or copy the skill folder (.agents/skills/cosa-platforms in coreos/coreos-assembler) into .claude/skills/cosa-platforms in your project. Claude Code loads it when a task matches its description.
How do I install Cosa Platforms in Codex?
Run `npx skills add coreos/coreos-assembler --skill cosa-platforms -a codex`. Or copy the skill folder (.agents/skills/cosa-platforms in coreos/coreos-assembler) into .agents/skills/cosa-platforms in your project. Codex loads it when a task matches its description.
Can I use Cosa Platforms in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add coreos/coreos-assembler --skill cosa-platforms -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/cosa-platforms, .gemini/skills/cosa-platforms, .github/skills/cosa-platforms and .opencode/skills/cosa-platforms in your project.
What does Cosa Platforms need to run?
SKILL.md names no scripts, command-line tools or credentials: Cosa Platforms is instructions for the agent only.
Does Cosa Platforms access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Cosa Platforms safe to install?
Our automated static check of SKILL.md flagged 2 warning(s): mentions a credentials file (ssh keys, cloud or package-manager tokens). Read the flagged lines before installing; the check is not a guarantee either way.
What licence does Cosa Platforms use?
Cosa Platforms is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Cosa Platforms use?
About 1.6k tokens (SKILL.md is roughly 6.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Cosa Platforms?
Skills that share tags, products or a category with Cosa Platforms: Cloud Cost Optimization (wshobson/agents, 40k stars), Terravision Cloud Diagrams (patrickchugh/terravision, 1.6k stars), Thesvg (glincker/thesvg, 2.8k stars) and Terrashark (LukasNiessen/terrashark, 714 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Cosa Platforms?
coreos (a GitHub organization) maintains it in coreos/coreos-assembler, which has 395 GitHub stars. The repository holds 4 skills in this directory. The repository was last updated on October 7, 2026.
Source: coreos/coreos-assembler on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.
