Detection reference · Credentials and secrets · Updated 2026-11-09

Anthropic API keys (sk-ant-)

Q: What does an Anthropic API key look like?

It starts with sk-ant-. The example in Anthropic's own docs is sk-ant-api03-... and the official SDKs read the key from the ANTHROPIC_API_KEY environment variable.

Q: Where should I look for a leaked Anthropic key?

Start with the name ANTHROPIC_API_KEY in pasted shell sessions, .env files and CI logs, because that is the variable the official SDKs read.

Anthropic API keys carry the sk-ant- prefix, which makes them one of the few sk- keys that can be attributed to a provider from the value alone.

At a glance

Reported as	API key or token
Category	Credentials and secrets
Severity	critical: direct account or system compromise, as with keys, passwords and tokens
Replaced with	`[SECRET]`
Where it runs	The same engine runs in the Chrome extension (free checks, on your device), in the web check and the Windows app (on Redaxa's server; the text is not stored) and in public repository checks.
Engine rule	scanner.ts, line 205

What it looks like

sk-ant- followed by the rest of the key (Redaxa's rule needs at least 16 letters, digits, hyphens or underscores after sk-).

Example

You paste

ANTHROPIC_API_KEY=sk-ant-api03-RedaxaDocsSampleNotRealRedaxaDocsSampleN

Redaxa returns

ANTHROPIC_API_KEY=[SECRET]

Reported as	API key or token
Severity	critical: direct account or system compromise, as with keys, passwords and tokens
Replaced with	`[SECRET]`
Probable service	Anthropic / Claude
Credential type	API key
Evidence	Anthropic-specific prefix; ownership and permissions are unverified.
Recommended action	Revoke or rotate if genuine

All values on this page are synthetic or published specimens; none is a working credential or a real person's data.

Key format, rotation and exposure

What Anthropic documents. The key is a static sk-ant-api... secret (the docs' own example is sk-ant-api03-...), sent as Authorization: Bearer <key> or in the legacy x-api-key header. The official SDKs read it from the ANTHROPIC_API_KEY environment variable, which is why that name is the first thing to look for in a pasted shell session, .env file or CI log.
Three key types. Personal keys act as you, service account keys act as a service account, and workspace keys (now legacy) belong to the workspace alone. Personal and service account keys stop working when the user or service account is removed from the organization; a workspace key does not, so an old workspace key is the one most likely to outlive its owner.
Disable or delete. In the Claude Console, Disable is reversible (the key can be re-enabled); Delete is permanent and the key is archived. Anthropic's guidance is to disable or delete any key you suspect has leaked. Our advice: disable first if you need a minute to deploy the replacement, then delete it, so nobody can switch an exposed key back on by mistake.
Expiry is set once. When a key is created you pick a preset of 3 hours, 1 day, one week or 30 days, a custom duration, or Never; it cannot be changed afterwards, and a key that has expired returns a 401 authentication error and cannot be reactivated. A short expiry limits how long a leaked key stays useful, but Anthropic is explicit that it does not replace keeping keys in a secrets manager.
Keep it out of browsers. Anthropic tells developers to call the API from their own back end, because anyone who loads a web page can read a key that sits in that page's code. For production workloads on AWS, Google Cloud or Azure, Workload Identity Federation exchanges a short-lived token instead, so there is no sk-ant-api... string to leak.

In a repository check

The same example, committed to a public repository and run through the free repository check:

.env.production: API key or token, listed for review (severity critical). Credential pattern matched. Validity is not tested; review and rotate if genuine.
Same value in tests/fixtures/.env.production: Test or example value, counted but not listed. Real-looking keys in test files are still reported.

Values are always shown masked in repository reports. No key is ever tested against its provider.

What it does not flag

ANTHROPIC_API_KEY=sk-ant-api03-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx: not flagged as API key or token; flagged as Secret in a configuration value. A run of x after the prefix is a placeholder, so it is not reported as an Anthropic key. Because the value still contains digits (api03), the configuration-value rule flags it: a placeholder that is cheap to dismiss.

If it already leaked

Disable or delete the key in the Claude Console (API keys), create a replacement, and review usage and spend for the period the key was exposed.

What Redaxa adds to every finding of this kind: If genuine, revoke or rotate it at the issuing service, update the application, then remove the exposed value from code and review Git history. Deleting it from Git does not revoke access.

Official guidance: platform.claude.com.

Step by step: pasted a key into ChatGPT, or committed one to GitHub.

Common questions

What should I do if my Anthropic API key leaked?

Disable or delete the key in the Claude Console under API keys, create a replacement, and review usage and spend for the period the key was exposed. Disable can be reversed; Delete is permanent.

What does an Anthropic API key look like?

It starts with sk-ant-. The example in Anthropic's own docs is sk-ant-api03-... and the official SDKs read the key from the ANTHROPIC_API_KEY environment variable.

Where should I look for a leaked Anthropic key?

Start with the name ANTHROPIC_API_KEY in pasted shell sessions, .env files and CI logs, because that is the variable the official SDKs read.

Can I put an Anthropic API key in a web page?

No. Anthropic tells developers to call the API from their own back end, because anyone who loads a web page can read a key that sits in its code.

Does an expired Anthropic API key still work?

No. A key that has expired returns a 401 authentication error and cannot be reactivated.

Check before you paste

The free Chrome extension runs this check inside ChatGPT, Claude, Gemini and Perplexity before a message is sent, on your device and with no daily limit. On the web you get 5 free prompt checks a day with no account.

Add the Chrome extension Try a free web check

Worried it is already in your code? Check a public GitHub repository free (3 checks a day).

Sources

Redaxa engine: scanner.ts line 205, credential-context.ts line 31 (Anthropic / Claude). Every output on this page was produced by that code when the page was built.
Claude Docs: Authentication (key types, Disable and Delete, expiration)