45 judges that evaluate AI-generated code for security, cost, and quality with built-in AST.
C elevated
- capability exposure inferred + 35
- tool safety inferred + 17
- trust mitigators mixed − 8
inferred mixed
The A–E grade is our heuristic synthesis — a "review this" prompt, not a verdict. Each factor is tagged by what backs it: attested (a verifiable record), reported (a third party's claim), or inferred (our own heuristic, e.g. permissions). See methodology.
grade last moved 1w ago · see ecosystem CVEs →
- C · 56 → C · 44
- C · 44 → C · 56
- C · 39 → C · 44
- C · 41 → C · 39
Inferred surface — each links to servers holding it:
No known CVEs for this server.
- high dangerous code
committed secret: Anthropic key, Stripe live key, private key · dynamic exec: eval(), new Function() · credential logged in 5 file(s)
- medium dangerous code
env-secret-flows-to-network-js: A process environment value reaches a network call. Review whether it is a credential leaving the process; the ordinary API-wrapper shape (read a ke
- medium toxic flow (lethal trifecta)
lethal trifecta reachable across this server's tool + source surface: private-data access + untrusted-content ingestion + network exfil (a leg is proven only in the analyzed source)
prompt-surface
shipped agent-instruction files + hidden-content / dangerous-code findings —
quoted from the analyzed source
analyzed commit 8432a66 · analyzer v33 · 2w ago
skills & prompt files 5
- agent-rules KevinRabun-judges-8432a66/.github/instructions/adding-a-judge.instructions.md
- agent-rules KevinRabun-judges-8432a66/.github/instructions/judges.instructions.md
- agent-rules KevinRabun-judges-8432a66/.github/instructions/release.instructions.md
- agent-rules KevinRabun-judges-8432a66/.github/instructions/secure-coding.instructions.md
- agent-rules KevinRabun-judges-8432a66/.github/instructions/vscode-extension-debugging.instructions.md
danger signals41
- dynamic code execution eval() KevinRabun-judges-8432a66/content/snippets/javascript/js-02-eval-injection.js :2
return eval("`" + template + "`"); - dynamic code execution eval() KevinRabun-judges-8432a66/content/snippets/javascript/js-08-prototype-pollution.js :4
return eval(obj.action); - dynamic code execution eval() KevinRabun-judges-8432a66/content/snippets/typescript/ts-02-eval-injection.ts :2
return eval(expr); - dynamic code execution eval() KevinRabun-judges-8432a66/content/snippets/typescript/ts-05-path-traversal.ts :4
return eval(content); - dynamic code execution eval() KevinRabun-judges-8432a66/content/snippets/typescript/ts-07-xss-reflect.ts :4
eval(html); - dynamic code execution new Function() KevinRabun-judges-8432a66/content/snippets/typescript/ts-08-unsafe-deserialize.ts :2
const fn = new Function("return " + raw); - dynamic code execution eval() KevinRabun-judges-8432a66/examples/quickstart.ts :20
const result = eval("getUser(" + userId + ")"); // SQL injection + eval - dynamic code execution eval() KevinRabun-judges-8432a66/examples/sample-vulnerable-api.ts :115
const filter = eval(req.body.filterExpression); // codeql[js/code-injection] Intentional — demo file for judges - dynamic code execution eval() KevinRabun-judges-8432a66/src/commands/benchmark-advanced.ts :1122
const result = eval(req.body.expression); - dynamic code execution eval() KevinRabun-judges-8432a66/src/commands/benchmark-ai-agents.ts :718
return eval(expression); - dynamic code execution new Function() KevinRabun-judges-8432a66/src/commands/benchmark-ai-agents.ts :723
return new Function("data", \`with(data) { return \\\`\${template}\\\`; }\`)(data); - dynamic code execution eval() KevinRabun-judges-8432a66/src/commands/benchmark-expanded-2.ts :2648
$result = eval("return " . $formula . ";"); - dynamic code execution eval() KevinRabun-judges-8432a66/src/commands/benchmark-expanded.ts :282
$result = eval("return " . $formula . ";"); - dynamic code execution new Function() KevinRabun-judges-8432a66/src/commands/benchmark-expanded.ts :2582
const compute = new Function("return " + expression); - dynamic code execution eval() KevinRabun-judges-8432a66/src/commands/benchmark-languages.ts :462
result = eval(expression) - dynamic code execution eval() KevinRabun-judges-8432a66/src/commands/benchmark-quality-ops.ts :1277
const result = eval(req.body.expression); - dynamic code execution eval() KevinRabun-judges-8432a66/src/commands/benchmark-security-deep.ts :1501
const result = eval(expression); - dynamic code execution eval() KevinRabun-judges-8432a66/src/commands/benchmark.ts :298
const result = eval(req.body.expression); - dynamic code execution new Function() KevinRabun-judges-8432a66/src/commands/benchmark.ts :441
const config = new Function("return " + req.body.config)(); - dynamic code execution eval() KevinRabun-judges-8432a66/src/commands/encoding-safety.ts :262
Checks: eval(), unsafe deserialization (pickle, marshal, yaml.load), SQL/HTML/XML interpolation, - dynamic code execution eval() KevinRabun-judges-8432a66/src/judges/ai-code-safety.ts :57
- No eval(), exec(), or dynamic code generation from untrusted input - dynamic code execution eval() KevinRabun-judges-8432a66/tests/all-judges-coverage.test.ts :24
eval(req.body.expr); - dynamic code execution eval() KevinRabun-judges-8432a66/tests/ast-analysis.test.ts :315
const result = eval(expr); - dynamic code execution eval() KevinRabun-judges-8432a66/tests/judges.test.ts :637
result = eval(user_input) - dynamic code execution eval() KevinRabun-judges-8432a66/tests/language-coverage.test.ts :255
return eval(data); - dynamic code execution eval() KevinRabun-judges-8432a66/tests/p0-p1-features.test.ts :196
eval(password); - dynamic code execution eval() KevinRabun-judges-8432a66/tests/patches-ext.test.ts :35
const result = eval(userCode); - dynamic code execution eval() KevinRabun-judges-8432a66/tests/pipeline-options.test.ts :315
eval(data); - dynamic code execution eval() KevinRabun-judges-8432a66/tests/subsystems.test.ts :8101
eval(userInput); - suspicious endpoint 169.254.169.254 (cloud metadata) KevinRabun-judges-8432a66/tests/cybersecurity.test.ts :664
const code = `fetch("http://169.254.169.254/latest/meta-data/iam/credentials");`; - suspicious endpoint 169.254.169.254 (cloud metadata) KevinRabun-judges-8432a66/tests/judges.test.ts :2282
const response = await fetch("http://169.254.169.254/latest/meta-data/iam/security-credentials/"); - suspicious endpoint 169.254.169.254 (cloud metadata) KevinRabun-judges-8432a66/tests/multi-judge-coverage.test.ts :179
const code = `const creds = await fetch("http://169.254.169.254/latest/meta-data/iam/credentials");`; - suspicious endpoint 169.254.169.254 (cloud metadata) KevinRabun-judges-8432a66/tests/subsystems.test.ts :646
' fetch("http://169.254.169.254/latest/meta-data");', - committed secret Anthropic key KevinRabun-judges-8432a66/src/commands/benchmark-ai-agents.ts :3946
sk-ant…(56 chars, redacted) - committed secret Stripe live key KevinRabun-judges-8432a66/src/commands/benchmark-infrastructure.ts :1150
sk_liv…(31 chars, redacted) - committed secret private key KevinRabun-judges-8432a66/tests/github-app-ext.test.ts :26
PEM private key block (redacted) - credential in logs credential in log KevinRabun-judges-8432a66/src/commands/benchmark-advanced.ts :557
logger.debug("Auth attempt", { email, password }); - credential in logs credential in log KevinRabun-judges-8432a66/src/commands/benchmark-expanded-2.ts :1642
console.log('New user registration:', { name, email, ssn, creditCard, password }); - credential in logs credential in log KevinRabun-judges-8432a66/src/commands/benchmark.ts :1027
logger.info("Login attempt", { username, password }); - credential in logs credential in log KevinRabun-judges-8432a66/tests/evaluator-deep2.test.ts :346
logger.info("User login", { email: user.email, ssn: user.ssn, password: req.body.password }); - credential in logs credential in log KevinRabun-judges-8432a66/tests/judges.test.ts :2772
console.log("User auth:", password, token, secret);
[](https://mcpobservatory.com/servers/github:KevinRabun/judges/security) Heuristic, inferred signals — false positives (legitimately powerful tools, forks, language ports) are expected. Treat each as "review this", not a verdict. See the ecosystem-wide picture on the security hub, or the fleet security of KevinRabun.