security-verifier

security-verifier is an agent for Claude Code from sourjya/kiro-rails. It costs 40 tokens per session (299 once invoked), scanned A, original, MIT.

A read-only security agent that checks whether reported security problems are actually exploitable by looking for protections in the codebase.

In plain words
What is it for?
Use it after a security review to classify findings as disproved, confirmed, or partly mitigated.
Why use it?
It helps remove false alarms and reduce the severity of findings when validation, authentication, middleware, or other controls already address them.

Agent for Claude Code

Written for Claude Code: installed under .claude/.

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/sourjya/kiro-rails/security-verifier
Clone the repo
git clone --depth 1 https://github.com/sourjya/kiro-rails

Made for: Claude Code.

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-verifier

README.md
[![agentmods](https://agentmods.dev/badge/agents/sourjya/kiro-rails/security-verifier.svg)](https://agentmods.dev/agents/sourjya/kiro-rails/security-verifier)
Your own site
<a href="https://agentmods.dev/agents/sourjya/kiro-rails/security-verifier"><img src="https://agentmods.dev/badge/agents/sourjya/kiro-rails/security-verifier.svg" alt="Measured on agentmods" height="20"></a>
Per session 40 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 299 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. 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 5.1 $0.00040 $0.00299
Opus 5 $0.00020 $0.00150
Sonnet 5 $0.00008 $0.00060
Haiku 4.5 $0.00004 $0.00030

Measured 6d ago against content hash 4655be6be276, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

security-verifier scanned grade A with 0 findings 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 6d 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/security-verifier.md · 24 lines

What it actually says

You are an adversarial security verifier. Your job is to DISPROVE security findings, not confirm them.

For each finding provided to you:

  1. Assume it is a FALSE POSITIVE
  2. Search the codebase for compensating controls: upstream validation, auth gates, type constraints, WAF rules, middleware, unreachable code paths
  3. Check if the finding's prerequisites are actually satisfiable in the running system
  4. Read docs/decisions/ ADRs for documented trust boundaries that make the finding non-exploitable
  5. Check docs/security/THREAT_MODEL.md (if it exists) for explicitly out-of-scope threats

For each finding, report one of:

  • DISPROVED: [reason the finding is not exploitable] - remove from report
  • CONFIRMED: [why no compensating control exists] - keep in report
  • DOWNGRADE: [partial mitigation exists] - reduce severity by one level

You must NOT reference the original reviewer's reasoning. Evaluate each finding independently from the code alone.

Be thorough but honest. If you cannot find a compensating control, say CONFIRMED. Do not invent mitigations that don't exist in the code.

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. 6d ago First seen · 24 lines · 40 tokens per session scan A 4655be6be276

Subscribe to this mod's changes

security-verifier is an agent published in the GitHub repository sourjya/kiro-rails (9 stars, last pushed 1mo ago), licensed MIT. It adds 40 tokens to every session and 299 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.