agentic-security

command

node

args

${CODEX_PLUGIN_ROOT}/scanner/bin/agentic-security-mcp.js --root ${CODEX_PROJECT_DIR:-${PWD}}

{
  "name": "agentic-security",
  "version": "0.75.1",
  "description": "Agentic-Security: SAST + SCA + secrets + IaC + LLM-sec scanner with deterministic fix toolchain. Exposes a Model Context Protocol (MCP) server with 17 tools that Codex CLI can call directly.",
  "author": {
    "name": "Ross Young",
    "email": "[email protected]"
  },
  "vendor": "Clear Capabilities Inc.",
  "homepage": "https://github.com/Clear-Capabilities/agentic-security",
  "license": "PolyForm-Internal-Use-1.0.0",
  "harness": "codex",
  "mcpServers": {
    "agentic-security": {
      "command": "node",
      "args": [
        "${CODEX_PLUGIN_ROOT}/scanner/bin/agentic-security-mcp.js",
        "--root",
        "${CODEX_PROJECT_DIR:-${PWD}}"
      ],
      "description": "agentic-security MCP server (17 tools: scan_diff, query_taint, explain_finding, find_rule_module, lookup_cve, synthesize_fix, verify_fix, apply_fix, append_scratchpad, read_scratchpad, append_agents_memory, read_agents_memory, synthesize_sca_upgrade, apply_sca_upgrade, query_triage_memory, query_findings_memory, query_cache_telemetry). Session root confined to the current project directory; apply_fix and apply_sca_upgrade require confirm:true and a valid last-scan.json HMAC signature; shadow findings are never auto-applied.",
      "trust": {
        "writes_files": true,
        "session_scope": "project_directory_only",
        "requires_confirm_for": [
          "apply_fix"
        ]
      }
    }
  },
  "notes": "Codex CLI support: validated against the protocol spec; not yet validated against a live Codex install. Report issues at https://github.com/Clear-Capabilities/agentic-security/issues."
}

{
  "name": "agentic-security",
  "version": "0.152.0",
  "description": "Agentic-Security is a powerful Claude Code plugin that automatically performs Application Security Testing (SAST, SCA, secrets detection, and more). Think of it as the easy button for making your Claude-generated code safe and secure.",
  "author": {
    "name": "Ross Young",
    "email": "[email protected]"
  },
  "homepage": "https://github.com/Clear-Capabilities/agentic-security",
  "license": "PolyForm-Internal-Use-1.0.0",
  "keywords": [
    "security",
    "sast",
    "sca",
    "secrets",
    "iac",
    "cve",
    "osv",
    "kev",
    "cisa-kev",
    "claude-code",
    "appsec",
    "prompt-injection",
    "llm-security",
    "ai-security",
    "mcp-security",
    "agent-security",
    "authz",
    "broken-access-control",
    "business-logic",
    "aspm",
    "llmsecops",
    "aibom",
    "ml-bom",
    "owasp-llm-top-10",
    "model-security",
    "sbom",
    "cyclonedx",
    "spdx",
    "pbom",
    "compliance",
    "owasp-asvs",
    "nist-ai-600-1",
    "scorecard",
    "red-team",
    "redteam",
    "promptfoo",
    "jailbreak",
    "llm-eval",
    "adversarial-testing",
    "ai-safety"
  ],
  "mcpServers": {
    "agentic-security": {
      "command": "node",
      "args": [
        "${CLAUDE_PLUGIN_ROOT}/scanner/bin/agentic-security-mcp.js",
        "--root",
        "${CLAUDE_PROJECT_DIR}"
      ],
      "description": "agentic-security MCP server (scan_diff, query_taint, explain_finding, find_rule_module, synthesize_fix, verify_fix, apply_fix). Session root is confined to the current project directory; apply_fix requires confirm:true and a valid last-scan.json HMAC signature; shadow findings are never auto-applied.",
      "trust": {
        "writes_files": true,
        "session_scope": "project_directory_only",
        "requires_confirm_for": [
          "apply_fix"
        ]
      }
    }
  }
}