soundcheck

Soundcheck

Automated OWASP security checks for Claude Code. 29 skills covering OWASP Web Top 10:2025 and OWASP LLM Top 10:2025 that auto-invoke when Claude writes vulnerable code patterns, rewrite the vulnerable section inline, explain the fix, and continue with your original task.

No configuration needed. No user intervention required.

---

Install

claude plugin marketplace add thejefflarson/soundcheck
claude plugin install soundcheck

After installation, all 29 skills are active in every Claude Code session. Claude will automatically invoke the relevant skill whenever it detects vulnerable code patterns.

Try it without installing (current session only):

claude --plugin-dir /path/to/soundcheck

---

How it works

Each skill has a description field that tells Claude when to invoke it. When you ask Claude to write code matching that description, Claude:

1. Invokes the skill automatically (visible in tool use)
2. Rewrites the vulnerable section with a secure alternative
3. Explains what was wrong and what the fix establishes
4. Continues with your original task

You do not need to ask Claude to check for security issues. Soundcheck runs in the background on every relevant code-writing task.

---

Trigger Reference

Code pattern	Skill invoked	OWASP
Authorization checks, resource ownership, IDOR, SSRF	broken-access-control	A01:2025
Server config, CORS, debug flags, security headers, secrets	security-misconfiguration	A02:2025
npm install, pip install, dependency manifests, CI/CD pipelines	supply-chain	A03:2025
Encryption, password hashing, random token generation, TLS config	cryptographic-failures	A04:2025
SQL queries, shell commands, templates with user input, eval, ORM raw queries	injection	A05:2025
Rate limiting, login flows, business logic, multi-step workflows	insecure-design	A06:2025
Login, sessions, JWT, password storage, MFA, API key management	authentication-failures	A07:2025
Deserialization, pickle/yaml load, software update verification, CI artifacts	integrity-failures	A08:2025
Logging, audit trails, error handlers that log, security event recording	logging-failures	A09:2025
Error handlers, try/catch, API error responses, exception propagation	exceptional-conditions	A10:2025
LLM prompt construction with user input, RAG pipelines, system prompts	prompt-injection	LLM01:2025
Rendering LLM output to UI, executing LLM-generated code, downstream LLM output use	insecure-output-handling	LLM02:2025
Fine-tuning pipelines, dataset ingestion, training data from external sources	training-data-poisoning	LLM03:2025
LLM input limits, inference backends, chatbot request handling, token budgets	model-dos	LLM04:2025
Loading pre-trained models, model registries, third-party LLM providers	llm-supply-chain	LLM05:2025
Sending PII/secrets to LLM, system prompts with sensitive data, LLM memory	sensitive-disclosure	LLM06:2025
LLM tool definitions, function schemas, plugin access controls	insecure-plugin-design	LLM07:2025
Autonomous agents, LLM-triggered write/delete/send actions, multi-step pipelines	excessive-agency	LLM08:2025
Displaying LLM output as fact, LLM-driven consequential decisions, no human review	overreliance	LLM09:2025
Inference API endpoints, model access controls, rate limiting on model serving	model-theft	LLM10:2025
MCP server definitions, tool schemas, tool handlers with file/shell/network access	mcp-security	LLM07:2025
OAuth2/OIDC flows, JWT validation, redirect URI handling, token endpoints	oauth-implementation	A07:2025
RAG pipelines, vector store ingestion, external document retrieval for LLM context	rag-security	LLM01:2025
Implementation plans for features, APIs, or components touching user data or auth	threat-model	A06:2025
Storing credentials/tokens/PII to local files, prefs stores, SQLite, or temp dirs	insecure-local-storage	A02:2025
URL scheme handlers, exported Android activities, IPC sockets, XPC service handlers	ipc-security	A01:2025
Agent-to-agent calls, subagent spawning, multi-agent pipelines	multi-agent-trust	LLM08:2025
User-supplied strings to LLM with Unicode control chars, homoglyphs, RTL override	token-smuggling	LLM01:2025

On-Demand

Command	What it does
/security-review	Full OWASP sweep — invokes all 29 skills, produces a severity-ranked findings report, rewrites Critical/High/Medium issues

---

GitHub Actions

Security Review Action

Runs a full Soundcheck security review against your repository, rewrites Critical/High/Medium findings, and opens a pull request with the changes and a severity-ranked findings table.