Identify control-flow, data-flow, and taint-driven security issues via Joern and CodeQL.
npx skills add https://github.com/alinaqi/maggy --skill cpg-analysis-alinaqi
Please help me install this Agent Skill. Skill: cpg-analysis Source: https://github.com/alinaqi/maggy/tree/main/skills/cpg-analysis Command: npx skills add https://github.com/alinaqi/maggy --skill cpg-analysis-alinaqi
It eliminates uncertainty in complex codebases by revealing control-flow paths, data-flow paths, and security-relevant taint propagation that are hard to infer from ASTs alone.
Use case example: Before a release, analyze how untrusted input travels through request handlers into authentication or database calls, then produce a targeted report of taint paths and likely vulnerability hotspots.
Use the cpg-analysis skill to run a CodeQL taint query for a suspected injection or unsafe redirect path in your target repository.
High-intent search queries and answers about installing and using this skill.
To trace taint propagation across functions, you can use interprocedural taint queries via CodeQL to map how untrusted input travels through request handlers into sensitive sinks like authentication or database calls.
Code property graph analysis merges AST, CFG, CDG, DDG, and PDG structures using Joern to map execution and dependency relationships across modules. You need it for deep security audits where basic syntax inspection cannot reveal complex control-flow or data-flow vulnerabilities.
Yes, you can use Joern and CodeQL together in a tiered workflow. Scope with initial tooling first, escalate to Joern for structural and flow queries, then apply CodeQL for security-focused, cross-function taint flows and vulnerability-style query results.
Yes, this approach supports complex refactoring impact analysis across large codebases by mapping execution and dependency relationships across functions and modules using the code property graph, revealing structural and flow changes that standard AST inspection misses.
To find injection or unsafe redirect paths, run a CodeQL taint query targeting the suspected vulnerability path in your target repository, which will produce a targeted report of taint paths and likely vulnerability hotspots.
Basic syntax inspection cannot identify control-flow paths, data-flow paths, and security-relevant taint propagation that span across functions and modules, making it insufficient for deep audits of authentication, payment, and request-handling code.
mindfold-aiGenerate project-specific Trellis coding specifications from repository architecture and source code patterns.
AgentSecOpsAnalyze Python code for security vulnerabilities and generate JSON reports.
ofelixdevGenerate custom Semgrep rules for security vulnerabilities using test-first development.
philo-grovesConstruct static analysis graphs for reachability, guard dominance, and path feasibility.
