Secrets Detection
Hunt (Secrets) finds leaked API keys, tokens, credentials and private keys in a repository and its git history, or in content you paste. It combines pattern matching with Shannon entropy analysis, and can check whether a detected credential is still active.
Secret types detected
Access keys and service account credentials for AWS, GCP and Azure.
Personal access tokens for GitHub and GitLab.
Keys for payment, messaging and email services.
Connection strings with embedded usernames and passwords.
Private key material and JWT signing secrets.
Shannon entropy analysis catches high-randomness strings that match no known pattern.
A standard scan covers the last 100 commits and a deep scan the full history, so secrets deleted from the current files are still found.
Optionally tests each detected credential against its provider and marks it active or inactive.
OpenAI (including project and service account keys), Anthropic, Hugging Face, Groq, Replicate, Perplexity, xAI and OpenRouter keys.
MCP configuration files (.mcp.json, .cursor/mcp.json, .vscode/mcp.json and others) with secrets in the file, plain-HTTP servers, unpinned packages, whole-disk file access or tools that run without approval.
Model files in a format that runs code when loaded, such as Python pickle and PyTorch checkpoints.
From repository to rotation
- 1
Choose your source
Give a Git repository URL and branch, or paste code, .env files, config or logs.
- 2
Pattern and entropy sweep
Known secret patterns are matched, and entropy analysis flags high-randomness strings that patterns miss.
- 3
Deduplication and verification
Duplicate findings are merged. With verification turned on, each detected credential is tested and marked active or inactive.
- 4
Findings and rotation guidance
Each finding shows the file, line and commit, with guidance on revoking and rotating that type of secret.
STRIPE_SECRET_KEY=sk_live_51H8...xK2
AWS_SECRET_ACCESS_KEY=wJalrXUtn...
DATABASE_URL=postgres://admin:p@ss@db/prodSpoorder Secrets Detection vs a pattern-only scanner
| Criterion | Spoorder | a pattern-only scanner |
|---|---|---|
| Detection | Patterns plus entropy analysis | Patterns only |
| Git history | Last 100 commits or full history | Current files only |
| Verification | Optional check that a credential is still active | Not included |
| Remediation | Rotation guidance per finding | Not provided |
Does Spoorder retain my source code after scanning?
No. A repository is cloned into a temporary directory for the scan and the clone is deleted afterwards. Pasted content is not stored. Spoorder keeps the finding metadata: file path, line number, secret type and a hash of the value.
Can secrets detection run on every push?
Not automatically today. Scans triggered by push or pull request webhooks do not run yet, so start scans from the dashboard.
What if a secret is in git history but removed from the current files?
A standard scan covers the last 100 commits and a deep scan covers the full history, for public and private repositories. Secrets found in an earlier commit are reported with that commit.
How does entropy analysis work?
Shannon entropy measures the randomness of a string. A high-entropy string that matches no known pattern, such as a 40-character alphanumeric value, is likely to be an API key or token and is flagged for review.
Does Spoorder check whether a secret is still active?
Only if you turn verification on for a repository scan. It makes a live call to the provider with each detected credential, so enable it only if you are authorised to test those credentials. For a private repository, only the files in which a secret was found are sent for that check, never the repository credential.
Start scanning
Start with a free health check. Secrets Detection is included from Builder ($69/mo).