Secret Detection ยท CVE & SBOM Module

The API key baked
into your container.

Secrets accidentally embedded in container images are one of the most common โ€” and most avoidable โ€” sources of credential exposure. Tacosec scans every layer of every image as part of your normal CVE workflow. No separate tool. No extra step.

How It Happens

Secrets end up in images more often than you think.

How it happens

Dev hard-codes DB password in Dockerfile

The result

Secret baked into every image layer ever built from it

How it happens

AWS key committed to .env, copied into build context

The result

Key live in container registry, accessible to anyone with pull access

How it happens

CI script echoes API token into image filesystem

The result

Credential extractable via docker inspect or layer extraction

Tacosec catches every one of these before the image leaves your build system.

Detection in Action

Found. Before it shipped.

tacosec scan payments-api:build-1491 --secrets

โ–ถ Scanning 6 layers for secrets and credentials...

โœ• SECRET FOUND AWS Access Key ID

Layer: sha256:a3f91c (stage: builder)

File: /app/.env.production

Match: AKIA[A-Z0-9]{16} โ€” verified AWS key pattern

โœ• SECRET FOUND GitHub Personal Access Token

Layer: sha256:b7d204 (stage: final)

File: /root/.gitconfig

Match: ghp_[A-Za-z0-9]{36} โ€” verified GitHub token

โœ• Build blocked โ€” 2 secrets detected ยท rotate credentials immediately

How It Works

Every layer. Every secret type.

Layer-by-layer image scanning

Secrets can hide in any layer โ€” the base image, an intermediate build stage, or the final layer. Tacosec scans every layer and pinpoints exactly which one contains the exposure.

100+ secret pattern detectors

From AWS access keys and GitHub tokens to generic high-entropy strings, Tacosec matches known secret formats and flags them before they ship.

Caught in CI โ€” before the registry

Secret detection runs as part of the same pipeline scan as CVE checking. A secret found in your PR never makes it to your container registry.

Verified vs. unverified findings

Not every high-entropy string is a real secret. Tacosec verifies patterns against known formats and reduces noise โ€” so your team investigates real exposures, not false alarms.

Detected Secret Types

We know what to look for.

AWS Access Keys & Secret KeysGitHub & GitLab Personal Access TokensStripe, Twilio, SendGrid API KeysSSH Private KeysDatabase connection strings with credentialsJWT signing secretsDocker Hub tokensAzure Storage connection stringsGoogle Service Account JSON keysGeneric high-entropy tokens

Included in CVE & SBOM

No extra setup. Already scanning.

Secret detection across every layer of every scanned image
100+ secret type detectors: cloud, SaaS, database, crypto keys
CI/CD blocking: secrets fail the build before reaching the registry
Exact layer and file path of the exposure โ€” no guessing
Part of the CVE & SBOM module โ€” no separate tool to run
Historical record of secrets found, when, and in which image

Stop shipping credentials in your images.

Secret detection is included with every Tacosec CVE scan โ€” no separate configuration needed.

Run a Scan