Webplode Marketplace

A Claude Code plugin marketplace with curated skills for AI-powered workflows.

Installation

Add the marketplace

/plugin marketplace add webplode/webplode-marketplace

Install plugins

/plugin install webplode@webplode-marketplace    # Core tools (beads, triage, scanner, prompt-leverage)
/plugin install khuym@webplode-marketplace        # Multi-agent coordination skills

Browse available plugins

Run /plugin in Claude Code and go to the Discover tab to see all available skills.

After installing, run /reload-plugins to activate.

Plugins

Webplode

Original skills built by the Webplode team.

prompt-leverage

Strengthen a raw user prompt into an execution-ready instruction set. Automatically detects task type (coding, research, writing, review, planning, analysis), infers intensity level, and adds structured scaffolding for better execution.

Usage:

Invoke as a skill in Claude Code:

/webplode:prompt-leverage

Or use the Python script directly:

python plugins/webplode/skills/prompt-leverage/scripts/augment_prompt.py "your raw prompt here"

Framework blocks applied selectively:

• Objective, Context, Work Style, Tool Rules, Output Contract, Verification, Done Criteria

Output modes:

• Inline upgrade, Upgrade + rationale, Template extraction, Hook spec

beads-rust

Agent-first issue tracking with the br CLI (beads_rust). Manage local-first, git-friendly issues, dependencies, and task workflows directly from Claude Code.

Usage:

/webplode:beads-rust

Key capabilities:

• Initialize and manage beads workspaces (br init, br create, br close)
• Query actionable work (br ready), blocked issues (br blocked), and stale tasks (br stale)
• Dependency-aware planning with br dep and br graph
• Git-friendly sync with br sync --flush-only / --import-only
• Full CLI reference and workflow patterns included

beads-viewer

Graph-aware triage engine for beads projects using the bv CLI. Computes PageRank, betweenness, critical path, and other graph metrics to help decide what to work on next.

Usage:

/webplode:beads-viewer

Key capabilities:

• Unified triage with bv --robot-triage (the mega-command)
• Dependency-respecting execution plans with --robot-plan
• Priority misalignment detection, sprint burndowns, ETA forecasts
• Label health, cross-label flow, and attention analysis
• Alerts, suggestions, and baseline drift detection

ultra-bug-scanner

Static analysis meta-runner that catches 1000+ bug patterns across 8+ languages before they reach production.

Usage:

/webplode:ultra-bug-scanner

Key capabilities:

• 4-layer analysis: lexical, syntactic (AST), semantic, and heuristic
• Supports JS/TS, Python, Go, C/C++, Rust, Java, Ruby, Swift
• Multiple output formats: text, JSON, SARIF, TOON, JSONL
• CI integration with --ci --fail-on-warning
• Baseline comparison for regression detection
• Pre-commit quality gate: ubs $(git diff --name-only --cached)

---

Khuym (v2.2.1)

Author: hoangnb24 (source repository) License: MIT

Khuym is a validate-first, multi-agent coordination skill suite for agentic software delivery. It provides a complete workflow pipeline — from feature exploration through planning, validation, parallel execution, review, and learning capture. All Khuym skills are authored by hoangnb24 and ported into this marketplace with full attribution.

Workflow Skills (run in order)

Skill	Description
using-khuym	Bootstrap meta-skill. Load first on any khuym project. Routes to the right skill, supports go-mode (full-auto) and quick-mode.
exploring	Socratic dialogue to extract locked decisions from the user BEFORE research or planning begins.
gkg	Codebase intelligence — architecture snapshots, dependency graphs, module relationships.
planning	Research, synthesize, and decompose features into executable beads with risk maps.
validating	Critical gate between planning and execution. Verifies plan soundness across 8 dimensions.
swarming	Orchestrates parallel worker agents for feature execution with file reservation and conflict handling.
executing	Per-agent worker loop (Flywheel). Register, claim bead, reserve files, implement, verify, close, report.
reviewing	Post-execution quality verification. 5 parallel specialist reviewers, 3-level artifact checks, UAT.
compounding	Captures learnings from completed work to make future work easier.

Standalone Skills

Skill	Description
debugging	Systematic debugging for blocked workers, test failures, build errors, and runtime crashes.
dream	Dream-style consolidation pass over artifacts and learnings with ambiguity resolution.
prompt-leverage	Strengthen raw prompts into execution-ready instruction sets for AI agents.
book-sft-pipeline	Fine-tuning pipeline for books — ePub extraction, SFT dataset creation, style transfer, LoRA training.
writing-khuym-skills	Meta-skill for creating and pressure-testing new khuym skills.

Usage:

/khuym:using-khuym          # Start here — routes to the right skill
/khuym:exploring             # Begin feature exploration
/khuym:planning              # Decompose into executable plan
/khuym:swarming              # Parallel execution

License

MIT