- Home
- Skills
- DevOps & Cloud
- CircleCI Config Check
More screenshots
Works with the AI Agents you already use
CircleCI Config Check
This skill finds those and the other CircleCI mistakes below, and tells you how to fix each one.
$5
CircleCI Config Check
Example session with this skill installed
Check this project with circleci-config-check.
- Read your context and instructions
- Compiled the circleci config check
$ python3 scripts/check_circleci.py examples/risky
BLOCK .circleci/config.yml:7 C03 DEV_OR_VOLATILE_ORB
slack: circleci/slack@volatile
Why: An orb is referenced as @volatile or with a dev: version. CircleCI marks volatile "Not Recommended": it pulls the last published version on every run. Development versions are mutable, any member of the orb's organization can overwrite them, and they expire 90 days after publication, so jobs run code nobody reviewed and later fail when the version is gone.
Safer: Pin a published semantic version (namespace/orb@1.2.3) and update it on purpose. Keep dev versions to the orb's own test pipeline.
BLOCK .circleci/config.yml:9 C03 DEV_OR_VOLATILE_ORB
BLOCK .circleci/config.yml:25 C01 LITERAL_SECRET_IN_CONFIG
BLOCK .circleci/config.yml:33 C01 LITERAL_SECRET_IN_CONFIG
BLOCK .circleci/config.yml:50 C02 INSECURE_DOWNLOAD_EXECUTED
BLOCK .circleci/config.yml:61 C05 REMOVED_EXECUTOR_IMAGE
BLOCK .circleci/config.yml:62 C05 REMOVED_EXECUTOR_IMAGE
BLOCK .circleci/config.yml:85 C04 CONFIG_REJECTED_BY_CIRCLECI
REVIEW .circleci/config.yml:8 C11 ORB_NOT_PINNED
REVIEW .circleci/config.yml:26 C12 MUTABLE_IMAGE_TAG
REVIEW .circleci/config.yml:27 C15 DEPRECATED_EXECUTOR
REVIEW .circleci/config.yml:41 C16 DOWNLOAD_RUN_WITHOUT_CHECKSUM
REVIEW .circleci/config.yml:44 C13 UNTRUSTED_VALUE_IN_SHELL
REVIEW .circleci/config.yml:54 C15 DEPRECATED_EXECUTOR
REVIEW .circleci/config.yml:72 C14 SSH_KEYS_NOT_SCOPED
REVIEW .circleci/config.yml:73 C13 UNTRUSTED_VALUE_IN_SHELL
REVIEW .circleci/config.yml:89 C17 DEPLOY_CONTEXT_ON_ANY_BRANCH
REVIEW .circleci/config.yml:93 C18 RETIRED_PIPELINE_VALUE
(Each finding also has Why and Safer lines, as in the first one.)
Verdict: BLOCK. Checked 1 file, 1 with CircleCI config content.
This is a static check of file text. NOT FLAGGED is not a guarantee.
Connects securely to your tools. The creator never sees your data.
About this skill
Since 2026-09-21 CircleCI rejects version 2.0 configs, branch filters such as /^(?!main).*/ and parameters that are used but never declared, so the pipeline fails before any job starts. A Sentry token pasted under environment: is readable by everyone with access to the project, and a deploy job with an aws-production context that runs on every branch hands those credentials to anyone who can push one. This skill finds those and the other CircleCI mistakes below, and tells you how to fix each one.
Your agent runs a scanner on every YAML file under .circleci/ (config.yml, continuation configs for setup workflows and dynamic config, packed config fragments) or on a config file you name. It understands orbs (registry, URL and inline), executors, jobs, commands, workflows, contexts, approval jobs, filters, pipeline parameters and << >> values, YAML anchors, and maps run: | blocks back to their real line numbers. It knows $NAME references, env_var_name orb parameters, test database passwords on service containers (also when a command reuses them) and settings such as OIDC_TOKEN_AUDIENCE: sts.amazonaws.com, so they are not reported as leaks. Rules with a CircleCI deadline compare with today's date. It never runs a job, the CircleCI CLI or the network. It returns BLOCK, REVIEW, or NOT FLAGGED, with the file, line, reason, and a safer approach, and links the CircleCI page behind each rule that has one. Secrets are never printed.
What it checks
BLOCK
- C01: passwords, tokens and keys written in the config (environment values, docker auth, orb and job parameters, parameter defaults, curl -u, Authorization headers, docker login -p, a literal echoed into docker login --password-stdin, npm _authToken, passwords in URLs, GitHub, AWS, CircleCI, npm, Slack and other token shapes, private keys)
- C02: scripts or binaries downloaded over http:// or with TLS checks off and then run without a checksum
- C03: orbs referenced as @volatile or with a dev: version (except the orb an orb repository publishes, in its own test pipeline)
- C04: config CircleCI no longer compiles: version other than 2.1, regex filters with negative lookarounds, possessive quantifiers or backreferences, undeclared parameters, pipeline parameters in inline orbs, job names with colons or parentheses, branches on a job, resource_class inside machine, an empty config.yml, and from 2026-12-14 the C19 validator warnings
- C05: removed executors: Ubuntu 14.04/16.04 and removed dated 20.04/22.04 machine images, removed Xcode versions (16.2.0 and 26.2.0 from 2026-12-07), Intel macOS and M1/M2 resource classes
REVIEW
- C11: partner and community orbs pinned only to a major or minor version, and URL orbs on a branch instead of a commit
- C12: Docker images with no tag or latest, machine images on edge
- C13: branch names, tags, commit messages, pull request titles, webhook bodies and string pipeline parameters pasted into run commands (also through a job parameter)
- C14: add_ssh_keys without fingerprints
- C15: legacy circleci/ images, circleci/classic and Ubuntu 20.04 machine images, Xcode 16.2.0 and 26.2.0 before their removal on 2026-12-07, machine: true with no image
- C16: curl | bash, bash <(curl ...) and download-then-run over HTTPS without a checksum or signature check
- C17: deploy, release, publish or signing jobs with a context that run on every branch, with no filter, approval job or workflow condition (non-prod, preview, staging and dev deploys are not counted)
- C18: retired pipeline values: pipeline.trigger_parameters.circleci., pipeline.trigger_parameters.github_app., pipeline.trigger_source, $CIRCLE_COMPARE_URL
- C19: config validator warnings that block pipelines from 2026-12-14 (enum default not in its enum, run step without a command, when beside a step, workflow with both when and unless, invalid command name, empty requires, bare docker: or steps:, bad parallelism or checkout method, undeclared job or command arguments, templated job type); reported as C04 from that day
What's included
- The scanner (Python 3.9+, standard library only)
- A rules guide with the reason, safer approach and official CircleCI source for each rule (C13 has none: no CircleCI page covers it)
- A risky and a safe example (an online shop with web, iOS and AWS deploy jobs), and the real report the scanner produces on the risky one
Limits It reads each file as written. Settings kept outside the repository are not seen: project environment variables, context restrictions, passing secrets to forked pull requests, orb allow-lists, config policies and branch protection. Orb source is not fetched, so what an orb does with its parameters is not followed. The removed image, Xcode and resource-class lists follow CircleCI's changelog up to November 2026. A few official documentation snippets are reported on purpose (for example echo of << pipeline.git.branch >> or of a string pipeline parameter in a run step); C13 rests on how CircleCI writes << >> values into the config, as no CircleCI page warns about it. The risky example has no C19 finding, because C19 becomes C04 on 2026-12-14. NOT FLAGGED is not a guarantee.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for DevOps & Cloud.
Static check of CircleCI config text. Project and organization settings in the CircleCI app, context restrictions, orb source code and CircleCI Server specifics are not checked. NOT FLAGGED is not a guarantee.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 4 days ago
Permissions
Allowed Hosts
Runs one local Python script in the terminal. It reads your project's files, writes nothing, and makes no network calls. The example config includes clearly fake, inert tokens (not-a-real-... NOTREAL) for the scanner to find. Nothing in it is run.