OPL Prompts & AI Agents

Supercharge Claude Code with specialized AI agents and prompts for BSV blockchain development, project automation, and workflow optimization.

What This Repository Does

This repository provides:

• Specialized AI Agents - Expert sub-agents for specific tasks (design, security, documentation, content, payments, auth, etc.)
• Slash Commands - Instant automation tools for common workflows
• Automation Hooks - Background workflows to enhance development
• Powerful Prompts - Reusable templates for complex operations

Installation

Full Plugin (recommended - includes 28 agents, 48 skills, 10 commands, 5 hooks):

/plugin install bopen-tools@b-open-io

Skills Only (for other agentic frameworks — install individually):

bunx skills add b-open-io/bopen-tools --skill <skill-name>

bunx skills add b-open-io/bopen-tools --skill agent-auditor
bunx skills add b-open-io/bopen-tools --skill agent-decommissioning
bunx skills add b-open-io/bopen-tools --skill agent-onboarding
bunx skills add b-open-io/bopen-tools --skill benchmark-skills
bunx skills add b-open-io/bopen-tools --skill charting
bunx skills add b-open-io/bopen-tools --skill check-version
bunx skills add b-open-io/bopen-tools --skill cli-demo-gif
bunx skills add b-open-io/bopen-tools --skill code-audit-scripts
bunx skills add b-open-io/bopen-tools --skill confess
bunx skills add b-open-io/bopen-tools --skill create-next-project
bunx skills add b-open-io/bopen-tools --skill critique
bunx skills add b-open-io/bopen-tools --skill deploy-agent-team
bunx skills add b-open-io/bopen-tools --skill devops-scripts
bunx skills add b-open-io/bopen-tools --skill front-desk
bunx skills add b-open-io/bopen-tools --skill frontend-performance
bunx skills add b-open-io/bopen-tools --skill generative-ui
bunx skills add b-open-io/bopen-tools --skill geo-optimizer
bunx skills add b-open-io/bopen-tools --skill github-stars
bunx skills add b-open-io/bopen-tools --skill hammertime
bunx skills add b-open-io/bopen-tools --skill hook-manager
bunx skills add b-open-io/bopen-tools --skill humanize
bunx skills add b-open-io/bopen-tools --skill hunter-skeptic-referee
bunx skills add b-open-io/bopen-tools --skill linear-planning
bunx skills add b-open-io/bopen-tools --skill mcp-apps
bunx skills add b-open-io/bopen-tools --skill nextjs-upgrade
bunx skills add b-open-io/bopen-tools --skill notebooklm
bunx skills add b-open-io/bopen-tools --skill npm-publish
bunx skills add b-open-io/bopen-tools --skill perf-audit
bunx skills add b-open-io/bopen-tools --skill persona
bunx skills add b-open-io/bopen-tools --skill plaid-integration
bunx skills add b-open-io/bopen-tools --skill process-cleanup
bunx skills add b-open-io/bopen-tools --skill reinforce-skills
bunx skills add b-open-io/bopen-tools --skill remind
bunx skills add b-open-io/bopen-tools --skill runtime-context
bunx skills add b-open-io/bopen-tools --skill saas-launch-audit
bunx skills add b-open-io/bopen-tools --skill shaders
bunx skills add b-open-io/bopen-tools --skill skill-publish
bunx skills add b-open-io/bopen-tools --skill statusline-setup
bunx skills add b-open-io/bopen-tools --skill threejs-r3f
bunx skills add b-open-io/bopen-tools --skill ui-audio-theme
bunx skills add b-open-io/bopen-tools --skill voice-clone
bunx skills add b-open-io/bopen-tools --skill wait-for-ci
bunx skills add b-open-io/bopen-tools --skill wave-coordinator
bunx skills add b-open-io/bopen-tools --skill x-research
bunx skills add b-open-io/bopen-tools --skill x-tweet-fetch
bunx skills add b-open-io/bopen-tools --skill x-tweet-search
bunx skills add b-open-io/bopen-tools --skill x-user-lookup
bunx skills add b-open-io/bopen-tools --skill x-user-timeline

Specialized AI Agents

Our 28 expert agents enhance Claude Code with specialized knowledge. See agents/ for full details.

Development & Architecture

• 🔵 prompt-engineer — Zack — Slash commands, agent skills, YAML frontmatter, Claude Code config
• 🏗️ architecture-reviewer — Kayle — Large-scale system design, refactoring, multi-file analysis
• 🔴 code-auditor — Nyx — Security vulnerabilities, comprehensive code audits
• 🛡️ security-ops — Paul — Runtime security, dependency scanning, supply chain analysis, OWASP
• 🚀 optimizer — Torque — Runtime performance, bundle analysis, Core Web Vitals
• 🧪 tester — Iris — Testing strategies, evals, skill benchmarking, CI automation
• 🧹 consolidator — Steve — File organization, deduplication, naming conventions

Platform & Infrastructure

• 🟠 devops — Zoro — Vercel + Railway + Bun stack, CI/CD, security scanning
• 🟢 database — Idris — PostgreSQL, MySQL, MongoDB, Redis, SQLite, Turso, Convex
• 📱 mobile — Kira — React Native, Swift, Kotlin, Flutter
• 🔗 integration-expert — Maxim — API integrations, webhooks, third-party services
• 🟠 mcp — Orbit — MCP server setup, diagnostics, publishing to NPM, PostgreSQL/Redis/GitHub MCP
• ⚡ nextjs — Theo — Next.js, React 19, Turbopack, Bun, Biome

Specialized Domains

• 🔷 creative-developer — Kris — Three.js, R3F, shaders, physics, interactive 3D prototypes
• 🗺️ cartographer — Leaf — MapLibre, Mapbox, Leaflet, CesiumJS, geospatial data
• 💚 payments — Mina — Payment integrations, Plaid, financial operations
• 🤖 agent-builder — Rowan — AI agent systems, tool-calling, multi-agent orchestration
• 📊 data — Mr. Data Accumulator — Data processing, analytics, ETL pipelines
• 📣 marketer — Caal — CRO, copywriting, SEO, launch strategy, email sequences
• 🗂️ project-manager — Sage — Linear planning, issue tracking, project organization

Content & Communication

• 🟣 designer — Ridd — UI/UX, Tailwind, shadcn/ui, accessibility, dark mode
• 🔷 documentation-writer — Flow — READMEs, API docs, PRDs, guides
• 🎵 audio-specialist — Juniper — ElevenLabs audio, sound effects, music generation
• 🩷 researcher — Parker — Web research, docs, APIs, parallel research strategies
• 🎮 community-manager — Ordi — 1Sat Ordinals Discord bot, BSV community engagement

Organization & Operations

• 🏢 front-desk — Martha — Team directory, routing, service provider lookup
• 💼 executive-assistant — Tina — Google Workspace, scheduling, communications
• 🌐 account-manager — Kurt — Public-facing sales, visitor qualification, bOpen.io chat
• 🔴 satchmo-live — Satchmo — Persistent agent at satchmo.dev, BSV knowledge base

Usage: "Use the [agent-name] to [specific task]"

Skills

Skills are context-triggered capabilities. They activate automatically or can be invoked directly. Install individually with bunx skills add b-open-io/bopen-tools --skill <name>.

X/Twitter

Content & Media

Development & Quality

Agent & Plugin Management

Operations & DevOps

Integrations

Organization

Slash Commands

Commands use category subdirectories: /category:command or /command for root-level commands.

• /bug-hunt - Adversarial bug hunt with 3 isolated agents — supports path or branch diff mode
• /prime - Context warm-up — loads git state, plugin inventory, and project conventions
• /question - Read-only Q&A mode — answers questions about the codebase without making changes
• /diagnose - Fan out 3-5 agents to investigate a bug from every angle simultaneously
• /impact - Map the full blast radius before changing a file or function
• /review-wave - 4 specialized reviewers examine changes simultaneously (security, perf, correctness, style)
• /hammertime - HammerTime behavioral rules — status dashboard (no args) or create a rule from a description
• /hammertime:status - HammerTime status dashboard (alias for /hammertime with no args)
• /hammertime:manage - Interactive rule management — enable, disable, remove, view, test rules
• /utils:context - Generate repo context snapshot for agents

Automation Hooks

Hooks are distributed automatically with the plugin — no manual installation needed.

HammerTime Stop Hook

HammerTime is a behavioral guardrail system that runs on every assistant response. It catches rule violations using three-layer scored detection:

Score thresholds: 0 = pass, 1-4 = Haiku verification (~500ms, ~$0.001), 5+ = direct block.

Loop safety: Each rule has a max_iterations field (default: 3). The hook tracks blocks per session and auto-allows exit when the limit is hit. Counters reset on new sessions. Set 0 for unlimited. State stored in ~/.claude/hammertime/state.json.

Full-turn evaluation: Rules can opt into scoring ALL assistant messages since the user's last message (not just the final one). This catches violations in intermediate responses — e.g., the model dismisses an error mid-turn, then the final message just says "Done." Set "evaluate_full_turn": true on a rule to enable. The hook reads the session transcript JSONL backwards (last 2MB max).

Ships with a built-in project-owner rule that prevents dismissing errors as "pre-existing" (full-turn enabled).

Commands

Debugging

export HAMMERTIME_DEBUG=~/.claude/hammertime/debug.log

Debug log shows elapsed time, score breakdowns, transcript reads, and phase decisions:

[   1ms] LAST_MSG length: 2847 chars
[   3ms] TRANSCRIPT: found 7d5d184f-...jsonl
[   8ms] TRANSCRIPT: collected 3 assistant blocks, 4201 chars
[   8ms] FULL_TURN: scoring 1 rules against 4201 chars
[   9ms] SCORE: rule 'project-owner' score=7 (kw=3, intent=1, cluster=1)
[   9ms] BLOCK: score 7 >= 5, skipping Phase 2

Rules live at ~/.claude/hammertime/rules.json. See the hammertime skill for the full rule authoring guide.

Custom Statusline

Moved to Plugin: Statusline is now distributed as the claude-peacock plugin.

Installation

/plugin marketplace add b-open-io/claude-plugins
/plugin install claude-peacock@b-open-io

Auto-configures on first session with:

• Project tracking - Shows CWD (⌂) and last edited project (✎)
• Lint status - Error/warning counts
• Git branch - Branch name with dirty indicator (*)
• Clickable file paths - OSC 8 hyperlinks to open in your editor
• Peacock themes - 24-bit true color from VSCode settings

Configuration

No configuration needed - auto-detects code directory and editor!

Optional overrides:

export CODE_DIR="$HOME/custom/path"    # Override auto-detected code directory
export EDITOR_SCHEME="vscode"          # Override auto-detected editor

See the claude-peacock plugin for full documentation.

Repository Structure

prompts/
├── agents/                 # Specialized AI agents
├── commands/               # Slash commands
│   ├── bug-hunt.md         #   /bug-hunt
│   ├── hammertime.md       #   /hammertime
│   ├── hammertime/         #   /hammertime:status, /hammertime:manage
│   ├── docs/               #   /docs:* commands
│   └── utils/              #   /utils:* commands
├── hooks/                  # Automation hooks (copy to ~/.claude/hooks)
├── skills/                 # Agent skills (each has SKILL.md + evals/)
├── benchmarks/             # Benchmark results (latest.json)
├── scripts/                # CLI tools (benchmark.tsx)
├── references/             # Reference documentation
├── tsconfig.json           # JSX config for benchmark CLI
├── README.md
└── QUICKSTART.md

Skill Benchmarks

Every skill ships with evals that prove it works. Each eval runs twice — once with the skill injected, once as a bare baseline — and an LLM judge scores each assertion. The delta is the signal.

Live results: bopen.ai/benchmarks

Eval Format

Add evals alongside any skill at skills/<name>/evals/evals.json:

[
  {
    "id": "basic-usage",
    "prompt": "Write a short README for a CLI tool called 'greet'",
    "assertions": [
      {
        "id": "has-install-section",
        "text": "The output includes an installation section",
        "type": "qualitative"
      },
      {
        "id": "has-usage-section",
        "text": "The output includes a usage section with an example command",
        "type": "qualitative"
      }
    ]
  }
]

Running the Benchmark CLI

# Run all skills with evals
bun run benchmark

# Run a single skill

# Custom model or concurrency
bun run benchmark --model claude-opus-4-6 --concurrency 5

Results are written to benchmarks/latest.json and automatically published to bopen.ai after each CI run.

Resume support: Each eval result is cached by content hash (benchmarks/cache/). If a run is interrupted, restarting picks up where it left off — no tokens wasted.

Writing Evals for Your Skill

Invoke the benchmark-skills skill to get guided help:

"Use the benchmark-skills skill to help me write evals for my skill"

Or ask the tester agent directly:

"Have the tester agent write evals for the x-research skill and run the benchmark"

CI Integration

Benchmarks run locally using your existing Claude session — no API key needed. Commit benchmarks/latest.json alongside your skill changes and push. CI validates that results are present and warns if you changed a skill without updating its benchmarks. bopen.ai picks up the committed results via ISR.

Key Features

🚀 Instant Productivity

• Pre-built commands for common tasks
• Expert agents for specialized work
• Automation that works in the background

🔗 Ecosystem Integration

• Works with our entire BSV development stack
• Integrates with BigBlocks, Sigma Identity, and more
• Compatible with init-prism project generation

🛠️ Extensible

• Create custom commands with prompt-engineer
• Modify agents for your workflow
• Build new automation hooks

Advanced Usage

Working with Agents

Agents can be explicitly requested for specific tasks:

"Use the prompt-engineer agent to create a deployment command"
"Have the bitcoin-specialist review this transaction builder"
"Ask the design-specialist about component library best practices"

Custom Workflows

Create project-specific automation by combining:

1. Specialized agents for expertise
2. Slash commands for automation
3. Hooks for background tasks
4. Prompts for complex operations

Common Use Cases

Bug Hunting

/bug-hunt

Documentation

"Have the documentation-writer create a comprehensive README"

Agent Context

/utils:context

Recommended Permissions

Some agents use CLI tools that require permission. To avoid repeated prompts, add these to your ~/.claude/settings.json:

{
  "permissions": {
    "allow": [
      "Bash(agent-browser:*)",
      "Bash(curl:*)",
      "Bash(jq:*)"
    ]
  }
}

Or use /permissions to add them interactively.

Skill Limits & Configuration

Claude Code has a default 15,000 character budget for skill metadata. When you have many skills installed, some may be truncated from Claude's context.

Symptoms

• /skills shows fewer skills than expected
• Claude doesn't recognize skills you know are installed
• "75 of 107 skills" type messages

Fix: Increase the Budget

Add to your shell profile (~/.zshrc or ~/.bashrc):

export SLASH_COMMAND_TOOL_CHAR_BUDGET=30000

Then restart your terminal and Claude Code.

Check Current Status

Run /context to see token usage and which skills are being truncated.

Tips & Best Practices

1. Use agents for expertise - They have specialized knowledge
2. Slash commands for speed
3. Combine tools - Agents + commands = powerful workflows
4. Keep updated - Pull latest from this repo to get new agents/prompts

Need Help?

• New to Claude Code? See our Quick Start Guide
• Browse examples: Check the design/ and development/ directories

Contributing

When adding new content:

1. Commands go in commands/ (root-level) or commands/[category]/
2. Agents go in agents/
3. Hooks go in hooks/
4. Skills go in skills/
5. Use the prompt-engineer agent for creating commands
6. Test thoroughly before committing

Skill Provenance

Two layers track skill authorship and integrity:

• skills-lock.json — Vercel Labs format recording third-party skill sources, versions, and SHA256 content hashes. Ensures reproducible installs and detects tampering.
• .clawnet/ directories — On-chain Bitcoin attestation (B + MAP + AIP + BAP ATTEST) for skills we author. Provides cryptographic proof of authorship anchored to the BSV blockchain.