AgentHive Inc. · Agent security

Lock down what your bots can do.

For teams running AI agents with tool access: a free MCP server and security Check, plus a $49 Field Kit playbook. Least privilege, audit and deny-by-default tools instead of system-prompt promises.

Run free Bot Lock Check Field Kit $49 · Pro $149
Weighted research themes

From 20 educational AI-security sources (2026). Top risks we bake into Check & Playbook:

16/20Prompt injection
16/20Logging & audit
12/20Least privilege

Also weighted: excessive agency, tool abuse, MCP supply chain, secrets, exfil, guardrails, token delegation.
Product

Start free with the MCP server. Add the playbook when you need it.

Defense in depth for LLM apps, agents, MCP servers, and multi-agent workflows.

⌘

Bot Lock MCP — free

MIT-licensed local MCP server: Ed25519 agent identity, AES-256-GCM vault, signed hash-chained audit log, deny-by-default scope checks, and a kill switch. Source on GitHub.

✓

Bot Lock Check

Interactive self-assessment scored from weighted AI security themes. Works offline in your browser. Run the free Check.

▣

Field Kit

Dense playbook plus deny-by-default tools, MCP allowlist, secrets, and kill-switch YAML templates.

⬡

Pro

Everything in the Field Kit, for teams that want to back the free MCP server. No Pro-only files today.

Research-backed

Controls ranked by how often experts talk about them

We weight Bot Lock Check and the playbook toward the themes that show up most — so you fix what matters first.

ThemeWeightCoverage
Prompt injection / jailbreak16/20
Logging / monitoring / audit16/20
Least privilege / IAM12/20
Excessive agency / autonomy11/20
Agent tool abuse11/20
Supply chain / MCP / plugins11/20
Secrets management10/20
Data exfiltration / privacy9/20
Guardrails / I/O filtering9/20
Token delegation / OAuth9/20
Pricing

Simple tiers. Live checkout.

The MCP server and Check are free. Field Kit and Pro are one-time Stripe checkouts on the First Deploy account. After pay you land on the playbook.

Free
$0
  • Bot Lock MCP server (MIT, runs locally)
  • Bot Lock Check (browser, offline-capable)
  • Weighted theme scoring
  • Shareable risk snapshot
Start Check
Popular
Field Kit
$49 one-time
  • Full Bot Lock Playbook
  • Policy YAML: deny-by-default tools
  • MCP allowlist, secrets, kill-switch
  • 30-day adoption plan
Get Field Kit

Live Stripe checkout.

Pro
$149 one-time
  • Everything in Field Kit
  • Same playbook chapters: agent identity (NHI), vault issuance & audience binding, audit
  • Backs the free MIT MCP server
  • No Pro-only files yet
Get Pro

Live Stripe checkout.

Honest positioning: Bot Lock reduces agent risk when you apply layered controls. It does not guarantee zero successful injections or replace authorized red-team verification.

How it works

Three steps, in order.

1 · Free

Run the Check

Score your agent setup against the weighted themes above. Runs in your browser, works offline, and gives you a shareable risk snapshot. Run the free Check.

2 · Free

Install the MCP server

MIT-licensed and local: Ed25519 agent identity, AES-256-GCM vault, signed hash-chained audit log, deny-by-default scope checks and a kill switch. Source on GitHub.

3 · $49 one-time

Apply the Field Kit

The full playbook plus policy YAML for deny-by-default tools, MCP allowlist, secrets and kill switch, and a 30-day adoption plan. Read the free preview.

Playbook preview

Threat model → controls → verify

Dense, defense-only guidance: principles, checklists, anti-patterns, and a verification playbook for systems you own. Read the contents and chapter 1 free.

Read the preview
Contact

Ship safer agents

Questions on Field Kit, Pro, or an enterprise rollout? Email Daniel Graham at AgentHive Inc.

Email Daniel

AgentHive Inc·About Daniel Graham·Consult·First Deploy·AI Nexus 360·Infrastructure·Flick·JobProof·IndexMe·Bot Lock

Built by Daniel Graham, founder of AgentHive Inc · 25 years in enterprise IT and telecom · Palm Coast, FL · 320-335-6186 · daniel@agenthiveinc.com