Start a new pentest engagement against a target. AUTHORIZED TARGETS ONLY. This initiates reconnaissance and begins the automated assessment. All findings are stored and correlated in the findings database. Poll get_engagement_status(eng_id) for phase progress. The caller (LLM agent and the human ...
AI agents invoke start_engagement to trigger actions in Pentest Ai. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
This tool executes external security operations (reconnaissance, scanning) against specified targets with real-world consequences. While it includes an authorization gate ('AUTHORIZED TARGETS ONLY'), the tool itself performs active operations that can trigger alerts, disrupt services, or generate significant logs depending on target configuration.
From the tool's definition 'Start a new pentest engagement against a target' and 'initiates reconnaissance and begins the automated assessment' — the tool executes real penetration testing operations against network targets, triggering security scanning and data collection whose…
Documented attack patterns abuse exactly the kind of access start_engagement gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Pentest Ai, and nothing reaches the server without passing your rules. This is the rule we recommend for start_engagement:
{
"version": "1",
"default": "deny",
"tools": {
"start_engagement": {
"limits": [
{
"counter": "start_engagement_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}start_engagement stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Start a new pentest engagement against a target. AUTHORIZED TARGETS ONLY. This initiates reconnaissance and begins the automated assessment. All findings are stored and correlated in the findings database. Poll get_engagement_status(eng_id) for phase progress. The caller (LLM agent and the human operator behind it) MUST have written authorization to test the target. See pentestai.xyz/aup. Pass auth_profile to log into the target before scanning. Use the secure credential-resolver path (pentest-ai auth profile add) so secrets never enter the MCP/LLM payload. Without it, auth-gated bug classes (race conditions, mass assignment, type confusion, authenticated SQLi/XXE/IDOR) cannot be reached. Pass respect_rate_limits=True to honor HTTP 429 / Retry-After responses with exponential backoff (capped at 30s, 3 retries). Recommended for real bug-bounty targets behind WAFs; default off preserves today's behavior. Pass strict_scope=True to refuse any request whose host is outside the engagement target's host. Also disables redirect- following in primitives so a 302 to attacker.com cannot pull the scan off-target. Bug-bounty programs care a lot about scope discipline; default off preserves today's wide-open behavior. It is categorised as a Execute tool in the Pentest Ai MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Pentest Ai MCP server in PolicyLayer and add a rule for start_engagement: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Pentest Ai. Nothing to install.
start_engagement is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the start_engagement rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for start_engagement. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
start_engagement is provided by the Pentest Ai MCP server (0xsteph/pentest-ai). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 51 Pentest Ai tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
51 Pentest Ai tools catalogued and risk-classified — across an index of 42,500+ MCP servers.