client-runtime-intelligence

Extracts routes, secrets, and trust assumptions from JavaScript bundles and source maps.

7|4|Updated Jun 22, 2026
One-click install
npx skills add https://github.com/dbx0/skills --skill client-runtime-intelligence-dbx0
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: client-runtime-intelligence
Source: https://github.com/dbx0/skills/tree/main/skills/methodology/triage/client-runtime-intelligence
Command: npx skills add https://github.com/dbx0/skills --skill client-runtime-intelligence-dbx0

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Front-end bundles and source maps expose API routes, role checks, hidden parameters, and hardcoded credentials, but manually reading minified JavaScript is slow and regex-only secret scanning misses credentials returned from functions rather than assigned to variables. ## Core Features & Use Cases - Source Map Recovery: Harvests bundle URLs, validates .map files across the whole file (avoiding the prefix-check bug that silently discards large application maps), and reconstructs original first-party source while filtering vendor noise. - Two-Pass Secret Detection: Combines a 67-rule regex bank with a secret-named-identifier proximity pass and a Shannon entropy sweep to catch credentials that assignment-based patterns miss. - Trust-Mismatch Analysis: Converts recovered routes, feature flags, role checks, and hidden parameters into a manual test queue focused on actor, state, and helper-endpoint differences. - Use Case: During a web application assessment, recover 5,000+ original source files from deployed source maps, surface a live client secret the regex bank missed, and validate it against its issuer before reporting. ## Quick Start Analyze the target application's JavaScript bundles and source maps to build an endpoint map, recover hardcoded secrets, and generate a prioritized manual test queue.

Frequently Asked Questions about client-runtime-intelligence

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I recover original source code from JavaScript source maps?▼

Resolve the sourceMappingURL comment or probe <bundle>.js.map directly, then parse sources[] against sourcesContent[] and write each entry to disk. Validate maps by searching the whole file for "sources", since a prefix check silently rejects large application bundles.

Why do regex secret scanners miss hardcoded credentials in JavaScript?▼

Regex banks match assignment syntax, so a credential returned from a function like getClientSecret() produces no match. A second pass flagging secret-named identifiers plus a Shannon entropy sweep over quoted literals catches these cases.

How do I filter vendor files from recovered source map output?▼

Exclude node_modules, webpack/bootstrap, webpack/runtime, core-js, and regenerator-runtime paths. Also drop paths containing embedded http/https hosts, since some builds inline dependencies fetched from raw.githubusercontent.com that pass a node_modules-only filter.

Should I report every credential found in a JavaScript bundle?▼

No, validate each credential against its issuer first. Azure AD error codes like AADSTS700016 or AADSTS7000215 indicate dead secrets, and public-by-design values like MSAL client IDs, Firebase web config, and Sentry DSNs are not findings.

What are common false positives when scanning JavaScript for secrets?▼

Frequent false positives include base64 font data matching AWS key patterns, the jose library's PKCS#8 format-check string matching PRIVATE_KEY, CSS selectors like [type=password], and placeholder values. Apply noise filters only to generic rules, never to vendor-prefixed patterns.