Skip to main content
CounterAgent is organized into subcommands. The top-level entry point is counteragent.
counteragent --help

audit

Security scanner for MCP servers. Maps findings to the OWASP MCP Top 10.

audit scan

Scan an MCP server for security vulnerabilities.
counteragent audit scan [OPTIONS]
OptionRequiredDescription
--transportYesTransport type: stdio, sse, or streamable-http
--commandWhen --transport stdioServer command to spawn (e.g., "python my_server.py")
--urlWhen --transport sse or streamable-httpServer URL to connect to
--checksNoComma-separated scanner list (e.g., "injection,auth")
--outputNoOutput file path (default: results/scan.json)
--formatNoOutput format: json or sarif (default: json)
--verbose / -vNoEnable debug logging
counteragent audit scan \
  --transport stdio \
  --command "python my_server.py"
Run specific scanners only:
counteragent audit scan \
  --transport stdio \
  --command "python my_server.py" \
  --checks injection,auth \
  --output results/my-scan.json
Use --format sarif with audit scan or audit report to generate SARIF 2.1.0 reports compatible with GitHub Code Scanning.

audit enumerate

Enumerate MCP server capabilities without running security checks.
counteragent audit enumerate [OPTIONS]
OptionRequiredDescription
--transportYesTransport type: stdio, sse, or streamable-http
--commandWhen --transport stdioServer command to spawn
--urlWhen --transport sse or streamable-httpServer URL to connect to
counteragent audit enumerate \
  --transport stdio \
  --command "python my_server.py"
Lists the server’s name, protocol version, tools, resources, and prompts.

audit list-checks

List all available scanner modules and their OWASP mappings.
counteragent audit list-checks
No options. Prints a table of scanner modules with OWASP ID, description, and readiness status.

audit report

Generate a report from saved scan results.
counteragent audit report [OPTIONS]
OptionRequiredDescription
--inputYesPath to saved scan results JSON file
--formatNoReport format: html, json, or sarif (default: html)
--outputNoOutput file path (defaults to input path with new extension)
Report generation is coming soon. The command is defined but not yet fully implemented.

proxy

Interactive MCP traffic interceptor — “Burp Suite for MCP.” Sits between client and server, intercepting JSON-RPC messages for inspection, modification, and replay.

proxy start

Start the proxy with the interactive TUI.
counteragent proxy start [OPTIONS]
OptionRequiredDescription
--transportYesTransport type: stdio, sse, or streamable-http
--target-commandWhen --transport stdioServer command to proxy
--target-urlWhen --transport sse or streamable-httpServer URL to proxy
--interceptNoStart in intercept mode (default: off)
--listen-portNoLocal port for SSE/HTTP (default: 8888)
--session-fileNoAuto-save session to this file
counteragent proxy start \
  --transport stdio \
  --target-command "python my_server.py"

proxy replay

Replay a captured session against a live server.
counteragent proxy replay [OPTIONS]
OptionRequiredDescription
--session-fileYesPath to a saved session file
--target-commandYes*Server command for replay (stdio)
--target-urlNoServer URL for replay (not yet implemented)
--outputNoSave replay results to JSON
--timeoutNoPer-message response timeout in seconds (default: 10.0)
--no-handshakeNoSkip auto-handshake if session already includes initialize
counteragent proxy replay \
  --session-file session.json \
  --target-command "python my_server.py"

proxy export

Export a session to JSON.
counteragent proxy export [OPTIONS]
OptionRequiredDescription
--session-fileYesPath to a saved session file
--outputYesOutput file path
--output-formatNoExport format (default: json)
counteragent proxy export \
  --session-file session.json \
  --output report.json

proxy inspect

Print session contents to stdout (non-interactive).
counteragent proxy inspect [OPTIONS]
OptionRequiredDescription
--session-fileYesPath to a saved session file
--verbose / -vNoShow full JSON payloads
counteragent proxy inspect --session-file session.json -v

Future modules

counteragent inject --help   # Tool poisoning & prompt injection [Phase 2]
counteragent chain --help    # Multi-agent attack chains [Phase 3]