Skip to main content
Glama
env-secret-exposure-analyzer-mcp

env-secret-exposure-analyzer-mcp

by vola-trebla

πŸ” env-secret-exposure-analyzer-mcp

npm CI License: MIT

Your AI agent is one debug session away from leaking your secrets.

MCP server that scans your project for secret exposure risks β€” hardcoded API keys, unprotected .env files, and console.log calls that print credentials at runtime. Before your agent accidentally reads them out loud.


πŸ€” The problem

You ask your AI agent to debug a config issue. It reads src/config.ts. Inside:

console.log('Config loaded:', JSON.stringify(config));
console.log(process.env.DATABASE_PASSWORD);

The agent now has your database password in its context. It might log it, include it in a summary, or pass it to another tool. And your .env isn't in .gitignore, so the next git push will do the rest.

None of this requires the agent to be malicious. It just needs to be helpful.

env-secret-exposure-analyzer-mcp catches this before it happens. πŸ”


Related MCP server: GuardX

πŸ› οΈ Tools

scan_for_secrets

Scans source files, config files, and .env files for 20+ secret patterns. Returns file path, line number, severity, and a masked preview β€” never the full value.

Detects:

  • AWS access keys + secret keys

  • GitHub tokens (ghp_, gho_, ghs_)

  • Stripe secret/publishable keys + webhook secrets (whsec_)

  • Anthropic, OpenAI API keys

  • SendGrid (SG.xxx), Twilio auth token + account SID

  • Google API keys + OAuth client secrets (GOCSPX-)

  • Slack tokens (xox*)

  • Private keys (-----BEGIN ... PRIVATE KEY-----)

  • Database URLs with embedded credentials (postgres://user:pass@host)

  • JWT secrets, session secrets, encryption keys

  • Sentry DSN, Datadog API key

  • Generic hardcoded passwords, secrets, tokens

Secret Scan Results
  Project:       /project
  Files scanned: 24
  Findings:      5

  [CRITICAL] .env:3 β€” AWS Access Key
    Preview: AKIA****MPLE
  [CRITICAL] .env:7 β€” Database URL with password
    Preview: post****sswd
  [CRITICAL] src/auth.ts:12 β€” Hardcoded JWT secret
    Preview: my-s****ecret
  [HIGH] .env:14 β€” Hardcoded session secret
    Preview: sess****key!
  [MEDIUM] .env:28 β€” Sentry DSN
    Preview: http****7890

check_gitignore_coverage

Checks whether sensitive files (.env, .env.local, secrets.json, private keys, certificates) are covered by .gitignore. Flags files that could be accidentally committed.

Gitignore Coverage Check
  Project: /project

  βœ— .env β†’ Add to .gitignore: .env
  βœ— .env.local β†’ Add to .gitignore: .env.local
  βœ“ secrets.json

scan_for_log_leaks

Scans source files for console.log / logger calls that print process.env variables or objects with secret-sounding names at runtime. Catches the most common "it's just a debug line" mistakes.

Log Leak Scan
  Project:       /project
  Files scanned: 18
  Findings:      3

  [CRITICAL] src/config.ts:8
    console.log("Config loaded:", JSON.stringify(config));
  [HIGH] src/server.ts:42
    console.log(process.env.AWS_SECRET_ACCESS_KEY);
  [HIGH] src/db.ts:15
    logger.info({ password: dbConfig.password });

πŸ§ͺ What it looks like in practice

A realistic .env with 20 secrets β€” database URLs, AWS, Stripe, Twilio, SendGrid, Google OAuth, Sentry, JWT secrets, encryption keys. Before this MCP: an AI agent reads the file, has no idea what's sensitive, and proceeds to use those values in generated code or responses.

After one scan_for_secrets call: 16 findings, all categorized by severity, all previews masked. The agent knows exactly what's dangerous before it touches anything.


⚑ Setup

{
  "mcpServers": {
    "secret-scanner": {
      "command": "npx",
      "args": ["-y", "env-secret-exposure-analyzer-mcp"]
    }
  }
}

πŸš€ Usage

"Scan this project for any secrets or API keys that might be exposed. Check if .env files are in .gitignore, and look for any console.log calls that might be leaking credentials."

The agent runs all three tools in sequence and reports a full picture: what's hardcoded, what's not protected, what's being logged.

Works great alongside:


License

MIT

Available Tools

5 tools
check_gitignore_coverageA

Check whether sensitive files (.env, .env.local, secrets.json, etc.) are properly covered by .gitignore rules. Flags files that contain secrets but could be accidentally committed.

Inputs: projectPath (required).

scan_ci_workflowsA

Scans GitHub Actions (.github/workflows/.yml), CircleCI (.circleci/config.yml), and GitLab CI (.gitlab-ci.yml) workflow files for dangerous secret interpolation patterns. Detects ${{ secrets. }} and ${{ github.token }} used directly in run: steps (log_leak risk β€” GitHub Actions logs the expanded plaintext) and ${{ github.event.pull_request.* }} / issue / commit content interpolated in shell commands (injection risk β€” attacker-controlled input). Returns file, job, step, pattern_found, risk, and recommendation.

Inputs: repo_path (required).

scan_for_log_leaksA

Scan source files for console.log / logger calls that may print environment variables or secrets at runtime. Catches patterns like console.log(process.env.SECRET) or logger.info({ apiKey }) before they reach production logs.

Inputs: projectPath (required).

scan_for_secretsA

Scan a project directory for hardcoded secrets, API keys, tokens, and passwords. Detects patterns like AWS keys, GitHub tokens, Stripe keys, private keys, and generic high-entropy strings. Returns file path, line number, severity, and a masked preview.

Inputs: projectPath (required), extensions.

scan_git_historyA

Scan the git commit history of a repository for secrets that were ever committed, even if later deleted. Detects hardcoded API keys, tokens, passwords, and other credentials added in diff hunks across recent commits. Returns each finding with its commit hash, file, line, severity, a masked preview, entropy score, whether the secret is still present in the working tree, and whether rotation is required. Respects .gitleaksignore.

Inputs: repo_path (required), since_days, last_n_commits.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A4/5.0

    Scored across 5 tools

    Disambiguation5/5

    Each tool targets a distinctly different aspect of secret exposure: hardcoded secrets in files, .gitignore coverage, log statement leaks, CI workflow risky patterns, and git history. There is no overlap or ambiguity between them.

    Naming Consistency4/5

    The verbs are mostly consistent (scan_for_* vs scan_* vs check_*), with scan_ as a common prefix. The inconsistent use of 'for' and the odd one out 'check_' create minor deviation, but the pattern is still recognizable.

    Tool Count5/5

    Five tools is well-scoped for a secret exposure analyzer. Each tool covers a relevant and non-redundant aspect, and the number is appropriate for the domain.

    Completeness5/5

    The set covers the primary surface areas for secret exposure: current source code, gitignore protection, runtime log leakage, CI workflow secret handling, and historical git leaks. This provides comprehensive coverage for the stated purpose.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      C
      maintenance
      Scans code for exposed secrets, API keys, tokens, and credentials across 69 patterns covering cloud services, AI platforms, payment providers, authentication services, and databases.
      MIT
    • A
      license
      Not graded
      quality
      B
      maintenance
      A keyless, defensive code-security auditor that scans codebases for hardcoded secrets, audits dependencies for known CVEs, and checks passwords against breach data using k-anonymity.
      1
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Local-first security check for AI coding agents β€” finds hardcoded secrets, exposed .env files, git-history leaks and vulnerable dependencies (OSV), entirely on your machine. Ask your agent "is this safe to ship?" and get a Launch Readiness score with a fix for every finding.
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Detects leaked credentials in source code with tools to scan text, files, and directories for API keys, tokens, and private keys across 30+ providers.
      1
      MIT