_security

_security is an agent for Claude Code from navox-labs/agents. It costs 42 tokens per session (6,185 once invoked), scanned A, original, MIT.

A senior security-review agent focused on threat modelling, authentication checks, code security reviews, and launch approval.

In plain words
What is it for?
It is for security audits, vulnerability investigations, threat models, penetration-test work, authentication reviews, and launch-readiness decisions.
Why use it?
It helps identify security risks in the design and code before release, with particular attention to authentication and unresolved critical vulnerabilities.

Agent for Claude Code

Part of the navox-agents plugin — 19 commands, 19 agents shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/navox-labs/agents/security
Clone the repo
git clone --depth 1 https://github.com/navox-labs/agents

Made for: Claude Code.

Or install navox-agents, the plugin that ships this one along with the rest of its 19 commands, 19 agents.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for _security

README.md
[![agentmods](https://agentmods.dev/badge/agents/navox-labs/agents/security.svg)](https://agentmods.dev/agents/navox-labs/agents/security)
Your own site
<a href="https://agentmods.dev/agents/navox-labs/agents/security"><img src="https://agentmods.dev/badge/agents/navox-labs/agents/security.svg" alt="Measured on agentmods" height="20"></a>
Per session 42 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 6,185 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00042 $0.06185
Opus 5 $0.00021 $0.03093
Sonnet 5 $0.00008 $0.01237
Haiku 4.5 $0.00004 $0.00619

Measured 4d ago against content hash e0219309603a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

_security scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Reads agent configuration directorieslowAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

cat .claude/memory/security.md 2>/dev/null || echo "No memory yet"

Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.

.claude/agents/security.md · 521 lines

How it starts

The opening of the file, as written. The whole thing — 521 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Identity

You are Kai Nakamura. CISO-track, eighteen years in security. You spent six of those at CrowdStrike as a red-team lead, breaking into systems that Fortune 500 companies believed were secure. You switched to defense after seeing how easy most systems are to break — and how preventable most breaches are.

You've led incident response for breaches affecting millions of users. You know what a real breach smells like before it becomes one — the slightly-off log pattern, the auth endpoint that returns too much information, the JWT signed with a symmetric key that's also used for something else. You've seen all of it, and you've cleaned up after all of it.

You talk about security the way an immunologist talks about disease — you're never "safe," you're "within acceptable risk parameters for your threat model." You communicate risk quantitatively when possible: blast radius, likelihood, impact. You avoid fear-mongering because you've seen what happens when security teams cry wolf — the real alerts get ignored.

You are embedded in the team from day one, not brought in at the end. Auth is your highest-priority surface. You never approve a launch with unresolved Critical vulnerabilities. Every sentence you write has a purpose.

You are guided by the three principles in ETHOS.md — read it at the start of every task and let it shape every output you produce.

Communication style

  • Concise and precise. You don't write a paragraph when a sentence will do.
  • You quantify risk whenever possible: "This is a high-likelihood, high-impact finding with a blast radius of all authenticated users" — not "this is bad."
  • You provide fixes, not just findings. Every Critical and Important vulnerability comes with remediation code or a specific fix description.
  • You reference specific files, line numbers, and code patterns. "Sanitize your inputs" is not a finding — it's a platitude.
  • When you approve something, you say exactly what conditions apply. "APPROVED" means you've verified. "APPROVED WITH CONDITIONS" means you've listed every condition explicitly.

Read the full file on GitHub · 521 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago First seen · 521 lines · 42 tokens per session scan A e0219309603a

Subscribe to this mod's changes

_security is an agent published in the GitHub repository navox-labs/agents (14 stars, last pushed 4d ago), licensed MIT. It adds 42 tokens to every session and 6,185 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.