codeql

Scans codebases for security vulnerabilities using CodeQL data flow and taint tracking analysis.

Updated Mar 22, 2026
One-click install
npx skills add https://github.com/TECH-HY/SKILLS --skill codeql-tech-hy
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codeql
Source: https://github.com/TECH-HY/SKILLS/tree/main/skills/codeql
Command: npx skills add https://github.com/TECH-HY/SKILLS --skill codeql-tech-hy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Running a thorough CodeQL security audit involves many failure-prone steps: building a high-quality database, modeling project-specific APIs, choosing the right query packs, and interpreting zero-finding results. This Skill orchestrates the entire pipeline so vulnerabilities are not silently missed due to bad databases, hidden suite filters, or missing data extensions. ## Core Features & Use Cases - Database Build Workflow: Detects the language, tries build methods in sequence (autobuild, custom command, multi-step tracing, macOS arm64 workarounds), and validates quality with baseline LoC and extractor error metrics. - Data Extension Generation: Enumerates sources and sinks CodeQL already recognizes, identifies gaps in project-specific wrappers (ORMs, shell runners, request parsers), and generates YAML source/sink/summary models. - Two Scan Modes: "run all" executes the security-and-quality suite across official, Trail of Bits, and Community packs; "important only" filters to high-precision security findings with a post-analysis severity threshold. - Use Case: Point it at a Python or C++ repository and ask for a full scan — it builds the database, creates data extensions for custom APIs, runs the analysis, and produces filtered SARIF results in a single output directory. ## Quick Start Ask the agent to run a CodeQL security scan on this repository and report the vulnerabilities found.

Frequently Asked Questions about codeql

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

FAQPage Schema
How do I run a CodeQL scan on my codebase?

Ask for a full scan and the skill builds a CodeQL database, generates data extensions for project-specific APIs, then runs the analysis. Results are written as SARIF files in an auto-incremented output directory like static_analysis_codeql_1.

What languages does CodeQL analysis support?

CodeQL supports Python, JavaScript/TypeScript, Go, Java/Kotlin, C/C++, C#, Ruby, and Swift. Interpreted languages need no build step, while compiled languages require build tracing via autobuild, custom commands, or multi-step tracing.

Why does my CodeQL scan return zero findings?

Zero findings usually indicate poor database quality, missing source/sink models, or silent suite filtering from passing pack names directly. Check baseline lines of code, extractor errors, and generate an explicit .qls suite file instead of relying on default pack suites.

How do I fix CodeQL build failures on macOS Apple Silicon?

Exit code 137 on Apple Silicon is an arm64e/arm64 mismatch between system tools and CodeQL's libtrace.dylib, not a real build failure. Use a Homebrew arm64 compiler with multi-step tracing or run the build under Rosetta x86_64 emulation.

When should I use Semgrep instead of CodeQL?

Use Semgrep for quick pattern searches, single-file checks, or when you cannot build compiled code. CodeQL is better for deep interprocedural taint tracking and comprehensive security audits where build capability exists.