⏳ This skill is pending AI review.
Scores will appear once the review pipeline completes.
cli-anything
Generate CLI harnesses for any software, making it agent-controllable. Use when asked to make a desktop app, GUI tool, or any software controllable via command line. Spawns an ACP Claude Code or Codex session to run the 7-phase build pipeline.
Choose how to use this skill
You do not need every option. Choose the path your AI client supports. The stable page stays the same; versioned files are immutable.
1. Native installer
This listing has no registered native installer command. Use the complete package or source fallback below, depending on what your client supports.
Do not guess an installer command or replace an existing version without reviewing the diff.
2. Complete package recommended
Download the ZIP when available. It includes SKILL.md plus the references, security notes and version metadata.
No complete ProSkills package is published for this listing yet.3. Prompt-only
Copy the prompt above when the agent can read the stable page or when you want to adopt the workflow without installing a skill.
Need only the instruction file?
Download SKILL.md only if your client requires a single file. The complete ZIP is safer for a full installation because it preserves the references and release context.
No path installs or executes anything by itself. Your agent still needs access to the project files. Before updating, compare the installed version and review the diff.
// RATINGS
Not yet listed on ClawHub or SkillsMP
// README
🧠 Agent Skills Pack
Battle-tested AI agent skills for OpenClaw. Built from 1,000+ hours of real-world autonomous agent orchestration.
These aren't demos. Every skill here has been forged through daily production use — handling real tasks, catching real bugs, surviving real compactions, and shipping real results.
⚡ Skills
🏗️ Execution & Quality
| Skill | What It Does | Lines |
|---|---|---|
| coding-agent | Full coding pipeline: Design → Plan → Execute → Review → Verify. Delegates to sub-agents with spec compliance + code quality gates. | 535 |
| convergent-review | Research → Design → Adversarial Review → Lock → Execute → Verify loop. Forces you to attack your own plan before building. | 475 |
| systematic-debugging | Structured debugging methodology. Use before proposing fixes, not after. Prevents the "bandaid instinct." | 261 |
| cli-anything | Generate CLI harnesses for any software — makes desktop apps and GUI tools agent-controllable. | 103 |
🔄 Autonomy & Self-Improvement
| Skill | What It Does | Lines |
|---|---|---|
| proactive-agent | Transform agents from task-followers into proactive partners. WAL Protocol, Working Buffer, Autonomous Crons, and battle-tested patterns. | 632 |
| self-improving-agent | Captures learnings, errors, and corrections. Reviews past mistakes before major tasks. Continuous improvement as a system, not a suggestion. | 647 |
| context-recovery | Automatically recovers working context after session compaction. Works across Discord, Slack, Telegram, Signal. | 233 |
🔒 Security & Vetting
| Skill | What It Does | Lines |
|---|---|---|
| forge | 9-gate security pipeline for vetting external resources. LITE/STANDARD/FULL risk tiers. Trust nothing by default. | 167 |
| security-auditor | OWASP Top 10, auth flow review, CORS/CSP, secrets handling, input validation, XSS/SQLi prevention. | 399 |
🔍 Discovery & Monitoring
| Skill | What It Does | Lines |
|---|---|---|
| mission-control | Real-time web dashboard for system health — agents, cron jobs, alerts, gateway status. | 293 |
| skill-discovery | Discover and safely acquire new skills from ClawHub. All acquisitions go through Forge. | 114 |
| find-skills | Helps users discover installable skills when they ask "can you do X?" | 133 |
📧 Integrations
| Skill | What It Does | Lines |
|---|---|---|
| outlook | Read, search, and manage Outlook emails + calendar via Microsoft Graph API. | 237 |
🚀 Installation
All skills at once
# Clone the repo
git clone https://github.com/aquilkurji/agent-skills-pack.git
# Copy skills to your OpenClaw installation
cp -r agent-skills-pack/skills/* ~/.openclaw/skills/
Individual skills
# Copy just the ones you want
cp -r agent-skills-pack/skills/coding-agent ~/.openclaw/skills/
cp -r agent-skills-pack/skills/convergent-review ~/.openclaw/skills/
Via ClawHub (coming soon)
clawhub install aquilkurji/coding-agent
🏛️ Philosophy
These skills encode hard-won lessons:
- "Should work" is banned. Every claim requires verification evidence.
- Review before you ship. The adversarial review step is NOT optional.
- Automation > documentation > rules. Scripts enforce what markdown suggests.
- Research before building. Stop reinventing wheels.
- Context is currency. Treat it like a finite resource because it is.
📋 Compatibility
- OpenClaw 2026.3.x+
- Models: Works with any model, optimized for Claude Opus/Sonnet
- Platforms: macOS, Linux (some skills require macOS for LaunchAgent support)
🤝 Contributing
Found a bug? Built an improvement? PRs welcome.
Each skill follows the AgentSkills spec:
SKILL.md— Main instruction file (mandatory)references/— Supporting documentationscripts/— Automation scripts
📄 License
MIT — use freely, build on top, make it better.
Built by @aquilkurji and Sherlock 🔍 — an autonomous AI agent running on OpenClaw.
// HOW IT'S BUILT
// FILES
2 files shipped with this skill. Inline text is the catalog copy. The upstream link is the source repository.
skills/cli-anything/SKILL.md3.3 KBupstream ↗
--- name: cli-anything description: "Generate CLI harnesses for any software, making it agent-controllable. Use when asked to make a desktop app, GUI tool, or any software controllable via command line. Spawns an ACP Claude Code or Codex session to run the 7-phase build pipeline." --- # CLI-Anything — Make Any Software Agent-Native > Source: ~/tools/cli-anything/ (GitHub: HKUDS/CLI-Anything) > Forge: PASSED 2026-03-14 ## What This Does Takes any software (GIMP, Blender, LibreOffice, OBS, any application) and automatically generates a full CLI harness that agents can use to control it. The generated CLI has: - Structured JSON output - Interactive REPL mode - Undo/redo support - Complete test suite - Auto-installs to PATH ## When to Use - User asks to "make X controllable" or "create a CLI for X" - User wants to automate a desktop application - User wants to control software that only has a GUI - Any request involving programmatic control of applications that don't have good CLI interfaces ## How to Execute **This is a heavy build task. Always spawn an ACP session — never run inline.** ### Option A: ACP Claude Code (recommended) ``` sessions_spawn({ task: "Use the cli-anything plugin to build a CLI harness for <software-path-or-repo>. Run: /cli-anything:cli-anything <path>", runtime: "acp", agentId: "claude", mode: "session", thread: true, cwd: "/path/to/software" }) ``` ### Option B: ACP Codex ``` sessions_spawn({ task: "Use CLI-Anything to build a harness for <software>. Follow the SKILL.md at ~/.codex/skills/cli-anything/", runtime: "acp", agentId: "codex", mode: "session", thread: true }) ``` ### Option C: Direct shell (for quick builds) ``` exec: cd ~/tools/cli-anything && claude --print "Use the cli-anything plugin. Build a CLI harness for <path>" ``` ## 7-Phase Pipeline The tool runs automatically through: 1. 🔍 **Analyze** — Scans source code, maps GUI actions to APIs 2. 📐 **Design** — Architects command groups, state model, output formats 3. 🔨 **Implement** — Builds Click CLI with REPL, JSON output, undo/redo 4. 📋 **Plan Tests** — Creates TEST.md with unit + E2E test plans 5. 🧪 **Write Tests** — Implements comprehensive test suite 6. 📝 **Document** — Updates TEST.md with results 7. 📦 **Publish** — Creates setup.py, installs to PATH ## After Build Generated CLIs are installed to PATH as `cli-anything-<softwarename>`. All agents on the machine can use them immediately: ```bash cli-anything-gimp --help cli-anything-gimp project new --width 1920 --height 1080 -o poster.json cli-anything-gimp --json layer add -n "Background" --type solid --color "#1a1a2e" ``` ## Refinement To expand an existing CLI's coverage: ``` /cli-anything:refine <path> # broad gap analysis /cli-anything:refine <path> "batch processing" # focused area ``` ## Pre-Built Examples The repo includes working harnesses for: Audacity, Blender, Draw.io, GIMP, Inkscape, Kdenlive, LibreOffice, OBS Studio, Shotcut, Zoom, AnyGen. These can be installed directly: ```bash cd ~/tools/cli-anything/<app>/agent-harness && pip install -e . ``` ## Install Locations - Source: `~/tools/cli-anything/` - Claude Code plugin: `~/.claude/plugins/cli-anything/` - Codex skill: `~/.codex/skills/cli-anything/` - OpenCode commands: `~/.config/opencode/commands/` - Methodology spec: HARNESS.md (in each location)README.mdupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
Plain-text copy, no login: /skill-bundles/aquilkurji-agent-skills-pack-skills-cli-anything.md