By robzilla1738
Offensive hotspot-driven hunt workflow for bugs, pain points, and inefficiencies — 13 analysis lenses, a zero-dependency toolkit (deterministic hotspots, structured findings, fingerprinted dedupe, baseline diffing, SARIF/HTML/markdown reports, CI exit codes), enforceable verification and coverage in CI, a fast Cursor-CLI fan-out variant (Composer 2.5 / Grok hunters, strong-model verify) and an optional E2B-backed Confirm rung, plus companion triage and fuzz skills. Report-only.
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Fast bughunt variant — fan out Composer 2.5 / Grok hunters via the Cursor CLI (cursor-agent), with the skeptic pass kept on a strong reasoner. Report-only.
Hunt for hidden bugs, pain points, and inefficiencies across a codebase and report ranked, reproducible, baseline-diffed findings (report-only)
Flush out hidden bugs dynamically with property-based / fuzz / differential harnesses
Triage a suspected bug or a set of findings — severity × confidence, minimal repro, standard report
Adversarial, hotspot-driven bug-hunting workflow for bugs, pain points, and inefficiencies across project types (iOS, macOS, web, services, terminal tools). A zero-dependency toolkit ranks hotspots, agents inspect risk areas through 13 analysis lenses, a mandatory skeptic pass refutes false positives, and strict merge gates can enforce verification and coverage in CI. Findings are fingerprinted, deduped, baseline-diffed, and rendered to markdown/HTML/SARIF with CI exit codes. Report-only by default. Use when the user invokes /bughunt, asks to find hidden bugs, audit code for defects, hunt pain points or inefficiencies, do a deep code review, gate CI on findings, or hunt for what tests miss.
Flush out hidden bugs dynamically by generating property-based tests, fuzz targets, or differential oracles, RUNNING them against the project's own test runner, then shrinking failures into regression tests or ready-to-commit harnesses. Use when the user invokes /fuzz, wants to harden a function, or when bughunt needs to confirm a Probable finding at runtime.
Triage a suspected bug or a set of findings into a ranked, evidence-backed report: assign severity and confidence, build a minimal reproduction or failing test, and emit the standard finding schema (markdown + JSON). Use when the user invokes /triage, hands you a crash or suspicious behavior to assess, or when bughunt needs to score, prove, baseline-diff, and report findings.
Provider-agnostic agent skills for AI coding assistants — packaged as a plugin marketplace that works with Claude Code, Cursor, and Codex.
Each skill is a folder of markdown files (SKILL.md + references) that teaches an agent how to
do one kind of work well: hunt bugs, review code, design native macOS/iOS UI, scaffold a
full-stack app. Plain files, no lock-in. Install as a Claude Code plugin in one command, or
copy the skill folder into Cursor/Codex.
Author: Robert Courson · robertcourson.com
| Plugin | Skills / commands | Description |
|---|---|---|
| bughunt-suite | /bughunt /triage /fuzz | Flagship. Evidence-first, hotspot-driven bug-hunting workflow for bugs, pain points, and inefficiencies — a zero-dependency toolkit ranks risk areas, agents sweep 13 lenses, strict merge gates enforce verification and coverage in CI, and findings are fingerprinted, baseline-diffed, and rendered to markdown/HTML/SARIF. Plus triage (severity × confidence + repro) and fuzz (runs property/fuzz harnesses). Report-only. |
| autoreview | /review | Hard acceptance gate before marking work complete. Reviews all session changes for production quality — correctness, architecture, maintainability, security, testing, cleanup. |
| scaffold | /scaffold | Product-driven greenfield Next.js — intake profiles, typed env, Vitest, tRPC, Neon/Drizzle, optional Clerk, Vercel. |
| macos-design | /macos-design | Design, critique, and scaffold native macOS apps — HIG, Liquid Glass, menus, toolbars, safe areas, accessibility, SwiftUI/AppKit. |
| macos-notch | /macos-notch | Design and implement macOS Dynamic Island / notch-style apps — NSPanel, geometry, state machine, module widgets. |
| macos-sandbox | /macos-sandbox | Smoke-test macOS .app/.pkg in disposable Tart VMs via macbox CLI or MCP. |
| web-marketing-landing | /landing-page | Design and scaffold a premium SaaS marketing home page (aurora hero, glass nav, bento grid) on Next.js + Tailwind v4. |
| svvarm | /svvarm | Opinionated design director for frontend interfaces. Reviews, fixes, and builds UI with a 38-pattern anti-slop standard — color, typography, layout, copy, interaction, production. Two-tier library: one ~5K-token digest for full builds, eleven deep references on demand. |
Portable replication guides (patterns captured from real projects) live in guides/ and are cataloged in INDEX.md.
/plugin marketplace add robzilla1738/roberts-skills
/plugin install bughunt-suite@roberts-skills
Install any other plugin the same way (/plugin install <name>@roberts-skills). Update with
/plugin marketplace update roberts-skills. Browse with /plugin.
Cursor loads skills from SKILL.md. Copy the skill folder(s) into a Cursor skills location:
git clone https://github.com/robzilla1738/roberts-skills.git
cp -R roberts-skills/plugins/bughunt-suite/skills/* ~/.cursor/skills/ # global
# or project scope: .cursor/skills/
Invoke with /<skill-name> (e.g. /bughunt) or attach with @<skill-name>.
Codex loads skills from .agents/skills/. Copy the skill folder(s) there:
cp -R roberts-skills/plugins/bughunt-suite/skills/* ~/.agents/skills/ # global
# or project scope: .agents/skills/ (read from CWD up to repo root)
Invoke with /skills or $<skill-name>.
.agents/skills/is read by both Cursor and Codex, so one copy there covers both tools.
Every skill works standalone — you don't need the whole repo or even the whole plugin. Grab just the skill folder you want:
# Claude Code (personal skills — available in every project)
npx degit robzilla1738/roberts-skills/plugins/bughunt-suite/skills/triage ~/.claude/skills/triage
# Cursor
npx degit robzilla1738/roberts-skills/plugins/bughunt-suite/skills/triage ~/.cursor/skills/triage
# Codex (also read by Cursor)
npx degit robzilla1738/roberts-skills/plugins/bughunt-suite/skills/triage ~/.agents/skills/triage
Swap the path for any skill listed in INDEX.md (plugins/<plugin>/skills/<name>).
Each skill folder is self-contained — SKILL.md plus its reference spokes.
Two caveats vs. a plugin install: you don't get the plugin's slash command registered
(invoke the skill by name instead), and updates are manual (re-run the command). Note
bughunt is the one skill best installed via its plugin — it reads its sibling skills
triage and fuzz as part of its pipeline.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimnpx claudepluginhub robzilla1738/roberts-skills --plugin bughunt-suitePlan and autonomously build a software task end-to-end. Recons the codebase, applies preloaded memory, decomposes into the right number of phases, gets one confirmation, then prepares a single ready-to-paste /goal command — one paste between you and done — that drives execution to completion with built-in retry, fix-spec recovery, and per-phase memory writeback. Works on Claude Code and Codex.
Opinionated design director for frontend interfaces. Reviews, fixes, and builds UI with a 38-pattern anti-slop standard — color, typography, layout, copy, interaction, production. One command, zero AI slop.
Hard acceptance gate: review all session changes for production quality before marking work complete.
Smoke-test macOS .app/.pkg in disposable Tart VMs via the macbox CLI or MCP.
Product-driven greenfield Next.js: intake profiles, typed env, Vitest, tRPC, Neon/Drizzle, optional Clerk, Vercel, intake add-ons (AI SDK, Resend, Stripe, Sentry, PostHog).
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Comprehensive PR review agents specializing in comments, tests, error handling, type design, code quality, and code simplification