C — Moderate risk signals Graded October 16, 2026Several concerns worth a closer look before you install.Based on public adoption and transparency signals as of October 16, 2026. Our opinion, not a code audit or security guarantee.

App connector · Developer tools, Cloud & infrastructure

fallow

Codebase intelligence for TypeScript and JavaScript: health, complexity, duplication, architecture, styling drift, and unused code from one graph. CLI, LSP, and MCP server. Zero config for over 100 frameworks.

Bottom line: Signal-based grade (2/7 points): heavily downloaded, no public source code found. This reflects adoption and transparency, not a code review.

Popularity score: 46/100
Rank: #192 of 500
Weekly downloads: 1.4M

What it does

Fallow analyzes TypeScript and JavaScript codebases to surface health metrics, complexity hotspots, code duplication, architecture violations, circular dependencies, design-system drift, and unused code through a unified dependency graph. It integrates across multiple workflows including CLI, GitHub Actions, LSP, VS Code extensions, and MCP servers.

Who it’s for

Engineering leaders, architects, and builders who need automated codebase quality visibility without manual audits. Teams using TypeScript/JavaScript who want to catch technical debt and architectural issues early in CI/CD pipelines.

Copy-paste setup

Check these steps against the project’s own instructions before running anything.

Via npm:

bash

npm install -g fallow
fallow analyze

As MCP server (Claude Desktop): Add to ~/Library/Application\ Support/Claude/claude_desktop_config.json (macOS) or Windows equivalent:

json

{
  "mcpServers": {
    "fallow": {
      "command": "npx",
      "args": ["fallow"]
    }
  }
}

Verify exact MCP configuration against the repository README (opens external site) before deploying.

Risks

Codebase exposure: Fallow analyzes your entire codebase structure and dependencies; if run via cloud MCP, architecture details could be transmitted externally unless using local-only mode.
No authentication isolation: The MCP connector does not natively support credential scoping; ensure it runs in isolated environments if analyzing sensitive repositories.
Supply-chain dependency: Depends on npm/Rust ecosystem; verify pinned versions and monitor for upstream vulnerabilities.
Prompt injection via code: Malicious code comments or variable names in your codebase could influence MCP responses if analysis results are fed directly into LLM prompts without sanitization.
LSP/editor access: VS Code and LSP integrations read local filesystem; misconfiguration could expose unintended directories.

How to stay secure

Least-privilege permissions

Run fallow in isolated CI/CD jobs with minimal filesystem access; restrict to repository root only, not system-wide directories.
Pin fallow version explicitly in package.json or Cargo.lock; never use latest or wildcard ranges for transitive dependencies.
If using VS Code LSP extension, disable it for untrusted workspaces; configure .vscode/settings.json to limit analysis scope to src/ and lib/ directories only.
Create a dedicated service account for MCP connector with read-only access to repositories; never share credentials with other tools.

Network isolation

Run fallow in local-only mode by default; disable cloud MCP until you fully understand data transmission paths.
Use an air-gapped or VPC-isolated environment for analyzing proprietary codebases; intercept and log all egress traffic from the MCP process.
If using GitHub Action, run within a self-hosted runner on private infrastructure; avoid GitHub-hosted runners for sensitive repositories.
Disable automatic telemetry/analytics; verify no metrics or codebase summaries are sent to upstream (Rust/npm registries or third-party servers).

Monitoring

Audit all files read by fallow at runtime; log to a centralized SIEM and alert on access to directories outside intended scope (.env, config, secrets).
Monitor outbound DNS/HTTPS from fallow process; block unexpected domains and flag any attempts to reach external services.
Review fallow's generated reports for information leakage (e.g., internal IP addresses, API endpoints, or hardcoded credentials in code paths); sanitize before sharing.
Check upstream dependency updates weekly; subscribe to security advisories for tree-sitter, swc, and core Rust crates fallow depends on.

Red flags to watch for

Slowdowns or timeouts during analysis on small codebases—may indicate exfiltration or malicious subprocess execution.
MCP responses incorporating code comments verbatim into LLM context without escaping—prompt injection risk; sanitize all code snippets before LLM processing.
Unexpected file access outside repository root (e.g., /home/, /tmp/, config directories)—sign of compromised or misconfigured MCP connector.
Dependency resolution failures or version mismatches after updates; fallow's supply chain is high-risk—verify Cargo/npm lockfiles are committed and reproducible builds work.

Alternatives

SonarQube/SonarCloud – Enterprise static analysis with broader language support and formal security scanning, but requires more infrastructure.
Knactualment – Lightweight dependency graph visualization, narrower scope than Fallow.
Archilyse / Architectural linters – Custom rule engines (ESLint plugins, TypeScript compiler APIs) for catching specific patterns.
Dependency-check / OWASP tools – Focus on supply-chain and CVE detection rather than code health metrics.

Sources

npmjs.com/package/fallow