Perform structured multi-model code review with an investigation phase.
npx skills add https://github.com/DSado88/squall --skill squall-deep-review
Please help me install this Agent Skill. Skill: squall-deep-review Source: https://github.com/DSado88/squall/tree/main/.claude/skills/squall-deep-review Command: npx skills add https://github.com/DSado88/squall --skill squall-deep-review
Complex codebases often hide bugs and security issues that single-model reviews miss. This Skill provides a rigorous, multi-model, investigation-first workflow to surface defects, guaranteeing a thorough, structured assessment before automated reviews.
Run a deep-review on your target codebase by initiating squall-deep-review with the appropriate models and targets.
High-intent search queries and answers about installing and using this skill.
A multi-model deep code review uses 3-5 different models with tailored prompts to investigate codebases, balancing coverage to reduce blind spots and surface complex security defects that single-model reviews miss.
To perform a deep code review, initiate a mandatory investigation phase to map control flow and form hypotheses, build per-model lenses, then invoke the review with deep enabled, passing investigation context and per-model system prompts.
Use an investigation-first code review for large codebases, complex systems, and critical security-sensitive code where thorough hypothesis generation and mapping control flow are required before running automated multi-model reviews.
Yes, you can orchestrate 3-5 models with per-model lenses tailored to investigation findings, synthesizing findings by consensus and cross-referencing them with initial hypotheses to prioritize fixes and reduce blind spots.
Single-model reviews often miss bugs and security issues hidden in complex codebases, whereas a multi-model approach applies diverse lenses to investigation findings, ensuring a structured assessment that prioritizes fixes by consensus.
ColdstartLabs-caInvestigate codebases to produce structured context summaries before addressing issues.
zachmayerOrchestrate parallel multi-model reviews across Codex, Gemini, and two Claude configurations.
binan-makerAggregate security scanner results into a prioritized summary with per-scanner status.
shaharKeisarAppsAnalyzes code quality, architecture, and documentation gaps in target repositories.
