skills.rest
escoffier-labs
escoffier-labsOFFICIAL
Explore all skills by this author →
WebsiteGitHub

bug-hunt

Scan codebases for confirmed correctness defects with adversarial verification and standardized reporting.

65|5|Updated May 13, 2026
Search Tags:#code review#verification#software testing#bug hunting#defect detection#false positive reduction#correctness sweep
One-click install
npx skills add https://github.com/escoffier-labs/brigade --skill bug-hunt-escoffier-labs
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: bug-hunt
Source: https://github.com/escoffier-labs/brigade/tree/main/registry/skills/bug-hunt
Command: npx skills add https://github.com/escoffier-labs/brigade --skill bug-hunt-escoffier-labs

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

AI-generated bug hunts often produce plausible but incorrect findings that waste developer time and erode trust in automated analysis. This Skill enforces mandatory adversarial verification to ensure only confirmed, actionable correctness defects are reported.

Core Features & Use Cases

  • Lens-based scanning: Systematically sweep code across 5 key defect categories (logic errors, error handling flaws, edge case failures, concurrency issues, and API misuse) to avoid missing common bug types.
  • Mandatory verification: Every candidate bug undergoes a skeptical adversarial review to refute unconfirmed findings, drastically reducing false positive rates.
  • Standardized reporting: Output follows a consistent schema with severity, effort, and triggering scenario details for easy triage and fix prioritization.
  • Use Case: An engineering team preparing for a production release can use this Skill to run a full correctness sweep of their codebase, get a prioritized list of confirmed bugs, and avoid wasting time on unverified hypotheses.

Quick Start

Use the bug-hunt skill to scan your repository's recently changed files for confirmed correctness bugs and generate a prioritized triage report.

Frequently Asked Questions about bug-hunt

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

FAQPage Schema
How do I reduce false positives during automated code review and bug hunting?▼

A correctness sweep systematically scans code across five key defect categories: logic errors, error handling flaws, edge case failures, concurrency issues, and API misuse. This lens-based scanning ensures broad coverage of common bug types while avoiding missed defects during pre-release reviews.

How do I prioritize bug fixes found during repository health checks?▼

Prioritize bug fixes by using a standardized reporting schema that scores every confirmed defect by severity and effort. This provides triggering scenario details for easy triage, helping engineering teams prioritize actionable correctness defects during production releases.

Can I run regression testing on any software project with accessible source code?▼

Yes, you can run regression testing and correctness sweeps across any software project with accessible source code. The process applies mandatory adversarial verification to identify confirmed bugs regardless of the specific framework or platform used.

What is the best way to verify plausible but incorrect AI-generated bug reports?▼

The best way to verify AI-generated bug reports is to apply a mandatory adversarial verification workflow that attempts to refute every unconfirmed finding. This eliminates plausible-but-wrong reports and ensures only verified correctness defects are delivered.

Why does automated defect detection sometimes report unverified code bugs?▼

Automated defect detection reports unverified bugs because AI-generated analysis often produces plausible but incorrect findings without strict validation. Applying an adversarial review process to refute unconfirmed candidate bugs solves this problem by ensuring only confirmed defects are reported.

Related Skills

Explore More Software Engineering Skills→
RedWoodOGRedWoodOG
174

simplify

Identify and fix code complexity issues in recently changed source files.

Community
Intermediate
Jia-Hong-PengJia-Hong-Peng

plan-devex-review

Audit developer experience gaps in APIs, CLIs, SDKs, and platforms.

Community
Advanced
ze-softwareze-software
50

ze-review-deep

Review Ze repository changes across security, concurrency, errors, tests, logic, data flow, API compatibility, rules, and docs.

Official
Advanced
dreadnodedreadnode
11

vuln-critic

Filter low-quality and false positive vulnerability findings from scanners and agent output.

Official
Intermediate