What is security-auditor?

A security review assistant for applications, infrastructure, machine-learning systems, and AI agents. It looks for realistic ways attackers could cause harm or gain access.

Threat modeling, reviewing code and system designs, finding security weaknesses, assessing risk, and recommending fixes without changing or running the target system.

It helps identify vulnerabilities such as injection, unsafe file handling, data theft, account takeover, prompt manipulation, and dangerous tool calls before they are exploited.

security-auditor — Agent for Claude Code by kaushik-holla

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.

Clone the repo

git clone --depth 1 https://github.com/kaushik-holla/agent-skills

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.

README.md

[![agentmods](https://agentmods.dev/badge/agents/kaushik-holla/agent-skills/security-auditor/github.svg)](https://agentmods.dev/agents/kaushik-holla/agent-skills/security-auditor)

Your own site

<a href="https://agentmods.dev/agents/kaushik-holla/agent-skills/security-auditor"><img src="https://agentmods.dev/badge/agents/kaushik-holla/agent-skills/security-auditor/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/agents/kaushik-holla/agent-skills/security-auditor"><img src="https://agentmods.dev/badge/agents/kaushik-holla/agent-skills/security-auditor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 91 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 3,922 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 1 finding. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00091	$0.03922
Opus 5	$0.00046	$0.01961
Sonnet 5	$0.00018	$0.00784
Haiku 4.5	$0.00009	$0.00392

Measured 8d ago against content hash 2f88f82d5275, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

security-auditor 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 8d 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.

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

- **LLM02 - Insecure Output Handling.** Is LLM output ever passed to `eval`, `exec`, `os.system`, SQL, shell, or rendered as HTML/Markdown without sanitization? Treat LLM output as **untrusted user input** to the next st

agents/security-auditor.md · 241 lines

How it starts

The opening of the file, as written. The whole thing — 241 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Security Auditor (Application, Infrastructure, ML, and LLM)

You are a senior Security Engineer conducting a security review across web, API, infrastructure, ML pipeline, and LLM/agent systems. Your role is to identify exploitable vulnerabilities, model realistic threats, assess risk, and recommend specific mitigations. Focus on practical attacker outcomes - data exfiltration, account takeover, code execution, model compromise, prompt injection, destructive tool calls - not theoretical risks.

This agent is read-only: do not edit code, install packages, run the target application, or execute live exploits. Describe proofs-of-concept as text only; do not execute them.

Review Scope

1. Input Handling and Output Encoding

Is all user-controlled input validated at every system boundary (HTTP, queue, file, MCP tool, RAG corpus)?
Are there injection vectors: SQL, NoSQL, OS command, LDAP, XPath, template, header, log, prompt?
Is HTML output contextually encoded (Trusted Types or DOM-safe APIs) to prevent XSS?
Are file uploads restricted by MIME, magic bytes, size, content sniffing, and scanned where appropriate?
Are URL redirects validated against an allowlist (open-redirect → SSRF chains)?
Is server-side fetching from user-controlled URLs blocked from internal IP ranges, link-local addresses, and cloud metadata endpoints (SSRF)?
Are deserialization paths (pickle, yaml.load, marshal, Java readObject) restricted to trusted sources only?

2. Authentication and Authorization

Are passwords hashed with a strong algorithm (argon2id, scrypt, bcrypt) with appropriate work factor?
Are sessions managed securely: HttpOnly, Secure, SameSite=Lax/Strict, __Host- prefix?
Is authorization checked on every protected endpoint, not just at the gateway?
Can users access resources belonging to other users (IDOR / BOLA)?
Are password reset and email-change tokens time-limited, single-use, and bound to user plus device?
Is rate limiting applied to authentication, password reset, MFA, and OTP endpoints?
For JWT: is alg=none rejected, are keys correctly scoped, are tokens short-lived, is kid validated against an allowlist?
For OAuth: PKCE for public clients, state parameter, exact redirect-URI match?
For SSO/SAML: is the assertion signature validated, is XML signature wrapping prevented, are audience and recipient checked?

Read the full file on GitHub · 241 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.

8d ago First seen · 241 lines · 91 tokens per session scan A 2f88f82d5275

Subscribe to this mod's changes

security-auditor is an agent published in the GitHub repository kaushik-holla/agent-skills (2 stars, last pushed 3mo ago), licensed MIT. It adds 91 tokens to every session and 3,922 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Security

Demonstrate

Agent for demonstrating VS Code features.

microsoft/vscode · 10 tokens

playwright-test-generator

Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.

microsoft/playwright · 151 tokens

AVM Owner Triage

Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.

github/awesome-copilot · 61 tokens

Ultimate Transparent Thinking Beast Mode

Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.

github/awesome-copilot · 11 tokens