AgentStack
MCP verified MIT Self-run

Decoy Tripwire

mcp-decoy-run-decoy-tripwire · by decoy-run

Security tripwires for AI agents. MCP tools that detect prompt injection, credential theft, and unauthorized tool calls in real time. Open source, zero dependencies, works with Claude Desktop, Cursor, Windsurf, VS Code, and Claude Code.

No reviews yet
0 installs
7 views
0.0% view→install

Install

$ agentstack add mcp-decoy-run-decoy-tripwire

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README — it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/mcp-decoy-run-decoy-tripwire)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
2mo ago

Declared compatibility

Claude CodeClaude DesktopCursorWindsurf

Compatibility is declared by the source manifest. End-to-end runtime verification is coming — see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps — measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Decoy Tripwire? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Decoy Tripwire

Know when your agents are compromised. Decoy Tripwire drops decoy MCP tools alongside your real ones — tools like execute_command, make_payment, access_credentials that no legitimate agent should ever call. When a prompt injection triggers one, the proxy pauses the compromised agent immediately and alerts you.

Every decoy returns a realistic error (timeout, permission denied). The agent sees a broken real tool, not a detection. Attackers don't know they've been caught.

Works with: Claude Desktop, Cursor, Windsurf, VS Code, Claude Code

🚀 Get Started

npx decoy-tripwire init

That's it. init signs you up, installs the local proxy, wraps your existing MCP servers, and drops the tripwires. Restart your MCP host — tripwires are live.

When a tripwire fires:

  • The compromised agent is paused for 10 minutes (auto-expires)
  • A desktop notification surfaces which tool was tripped
  • Every wrapped MCP server denies subsequent calls from that agent in sub-ms
  • Full context appears in your dashboard

Clear the pause early with npx decoy-tripwire resume .

🧑‍💻 Install

npx decoy-tripwire init                    # Sign up and install (wraps upstreams by default)
npx decoy-tripwire init --no-wrap          # Install without wrapping existing MCP servers
npx decoy-tripwire login --token=xxx       # Log in with an existing token

Requires Node.js 18+. Zero runtime dependencies.

🎓 Docs

🛠 Commands

# Monitor
npx decoy-tripwire test                    # Fire a test trigger
npx decoy-tripwire status                  # Local pauses + hosted triggers
npx decoy-tripwire watch                   # Live tail of triggers

# When a tripwire fires
npx decoy-tripwire resume        # Clear an auto-pause immediately
npx decoy-tripwire resume --all            # Clear every pause
npx decoy-tripwire lock          # Turn an auto-pause into a permanent block
npx decoy-tripwire lockdown on             # Any tripwire hit pauses every agent

# Manage
npx decoy-tripwire agents                  # List connected agents
npx decoy-tripwire config                  # View alert configuration
npx decoy-tripwire upgrade                 # Upgrade to Team (via dashboard)
npx decoy-tripwire uninstall --confirm     # Remove from all MCP hosts

All commands support --json for scripting and --token=xxx to override the stored token.

🪤 Tripwire Tools

| Tool | What it traps | Severity | |------|--------------|----------| | execute_command | Shell execution | Critical | | write_file | File system persistence | Critical | | make_payment | Unauthorized payments | Critical | | authorize_service | Trust grants to external services | Critical | | modify_dns | DNS record hijacking | Critical | | read_file | Credential theft | High | | http_request | Data exfiltration | High | | database_query | SQL execution | High | | access_credentials | API key theft | High | | send_email | Phishing via agent | High | | install_package | Supply chain attack | High | | get_environment_variables | Secret harvesting | High |

Plus dynamically generated tools from 6 threat categories (cloud infrastructure, secrets management, payments, CI/CD, identity, network). Each deployment gets a unique, deterministic set.

🧠 How auto-block works

init rewrites each MCP host config so upstream servers run through node proxy.mjs -- . The proxy intercepts every tools/call:

  1. Checks the shared pause registry at ~/.decoy/pause.json — if the agent is paused, denies immediately.
  2. If the call is a tripwire, returns a fake error and writes a 10-min pause entry for the agent.
  3. Otherwise forwards to upstream.

Every proxy instance reads the registry on its hot path, so one tripwire hit blocks every wrapped server in the same process lifecycle. Sub-ms. Works offline. Dashboard sync is fire-and-forget.

Turn on lockdown mode to escalate — any tripwire pauses every agent, not just the one that tripped.

📦 Plans

| | Free | Team ($29/user/mo) | Business ($99/user/mo) | |---|---|---|---| | Tripwires (12+ dynamic) | Yes | Yes | Yes | | Auto-block via local proxy | Yes | Yes | Yes | | Email alerts | Yes | Yes | Yes | | 7-day history | Yes | | | | SARIF/JSON export | Yes | Yes | Yes | | Slack/webhook alerts | | Yes | Yes | | Agent profiles + fingerprinting | | Yes | Yes | | 90-day history | | Yes | Yes | | Threat intel API | | Yes | Yes | | Security testing | | Yes | Yes | | OWASP compliance reports | | | Yes | | Custom detection rules | | | Yes | | Gateway integrations | | | Yes |

🚢 Release Notes

See [CHANGELOG.md](CHANGELOG.md) or the hosted changelog.

🤝 Contribute

See [CONTRIBUTING.md](CONTRIBUTING.md).

🔗 Related

📝 License

MIT — see [LICENSE](LICENSE).

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.