# I need to scan code for vulnerabilities

> Find vulnerabilities, leaked secrets and security holes in code, servers and apps.

The best fits to scan code for vulnerabilities on Agent Switchboard, as of 20 Nov 2026: SafeWeave, HexStrike AI and Semgrep, of 20 tools that do it.

## Quick answers

- **Which official tools scan code for vulnerabilities?** Snyk Agent Scan, Codex Security and deepsec, among 7 that are official.
- **Which have a hosted MCP server?** None of the 20 tools lists a hosted MCP server address.
- **Which need no API key or sign-in?** HexStrike AI, ToolHive and Nightcrawler, among 8 that need none.

## Best fits

1. [SafeWeave](https://agentswitchboard.dev/agents/safeweave) (mcp): SAST Scan: Run static analysis security testing on source code and surface vulnerabilities by severity level
2. [HexStrike AI](https://agentswitchboard.dev/agents/hexstrike-ai) (mcp, cli): Vulnerability Scanning: Runs web, network, and binary scanners against a scoped target and normalizes findings into structured results.
3. [Semgrep](https://agentswitchboard.dev/agents/semgrep) (api, cli): Scans source code with AI-assisted SAST, SCA, and secrets detection, finding 8x more true positives with 50% less noise than standalone foundation models.
4. [ToolHive](https://agentswitchboard.dev/agents/toolhive) (cli, api, mcp): Secrets Management: Stores encrypted secrets with `thv secret` and injects them into MCP server containers without exposing local credentials.
5. [Snyk Agent Scan](https://agentswitchboard.dev/agents/snyk-agent-scan) (cli · Official): Security scanner for MCP servers and agent skills. Detects prompt injections, tool poisoning, and sensitive data risks across Claude, Cursor, Windsurf, and 10+ other agents.
6. [Nightcrawler](https://agentswitchboard.dev/agents/nightcrawler) (cli): Local AI pentesting agent that runs reconnaissance and vulnerability probing from your own hardware without sending scope data to a cloud service.
7. [Shannon](https://agentswitchboard.dev/agents/shannon) (cli): Pentests web apps and APIs you are authorized to test by reading their source, running real exploits, and reporting only findings with a working proof of concept.
8. [1Password Environments MCP](https://agentswitchboard.dev/agents/1password-environments-mcp) (mcp): Add Secrets Without Exposure: Adds secrets to an Environment and lists their names, while the server never returns secret values to the agent.
9. [Codex Security](https://agentswitchboard.dev/agents/codex-security) (cli, api · Official): Finds, validates and patches code vulnerabilities in repositories and Git changes, from the cs CLI, a TypeScript SDK, a GitHub Action or Codex.
10. [deepsec](https://agentswitchboard.dev/agents/deepsec) (cli · Official): Scans large codebases for vulnerabilities by running coding agents in sandboxes, with their API keys kept outside the sandbox.
11. [PentAGI](https://agentswitchboard.dev/agents/pentagi) (api, cli): Pentest Reporting: Compiles findings, evidence, and remediation guidance into a structured report at the end of an engagement.
12. [Agent Vault](https://agentswitchboard.dev/agents/agent-vault) (cli): Credential Injection: Inject API keys and secrets at the network layer so agents never handle credentials directly

All 20, with filters: https://agentswitchboard.dev/browse?task=scan-for-vulnerabilities

## For agents

- This task as JSON: https://agentswitchboard.dev/api/tasks/scan-for-vulnerabilities
- One pick for your user's need, with the next step: https://agentswitchboard.dev/api/recommend?need=scan%20code%20for%20vulnerabilities (or the `recommend` MCP tool at https://agentswitchboard.dev/api/mcp)
- Every task: https://agentswitchboard.dev/tasks · the whole site: https://agentswitchboard.dev/llms.txt

Updated 20 Nov 2026 with the weekly drop. Ranked by how well each listing matches the task, never by payment. Page: https://agentswitchboard.dev/tasks/scan-for-vulnerabilities
