AgentForge

Guided authoring of OpenClaw skills and agents, with templates

Works with setup

Test report

Verdict
Works with setup
Score
7.6/10
Tested
Jul 24, 2026
Environment
Claude Code 2.x (agent harness)

This skill is no longer available upstream. Our re-check on Aug 10, 2026 couldn't find it any more (repo unreachable/deleted). The test below is what we measured on Jul 24, 2026 and we're leaving it up as a record — but there is nothing left to install, so we've removed the command.

A Russian-language meta-skill for authoring OpenClaw skills and agents. Following the body produced a tighter skill scaffold than a freeform baseline: it enforces the workflow/role/data-driven/hybrid type taxonomy, puts triggers in the description, and its pitfalls table flags real traps like the unsupported 'version' frontmatter field and keeping data in data/ not memory/. The larger agent-creation half depends on the OpenClaw runtime (openclaw.json, gateway, sessions_send, cron) and could not be exercised or verified here. Clean on security: the body and templates contain no exfiltration, covert instructions, or self-escalation.

Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works

  • Installs cleanly 5/5
  • Triggers reliably 4/5
  • Output vs. baseline 6/10
  • Docs & honesty 4/5

What AgentForge does

A meta-skill that walks you through creating, improving, and auditing OpenClaw skills and agents via three modes: build a skill (11 steps), build an agent (9 steps, with memory and self-improvement), or refine an existing one (5 steps). Triggers on phrases like 'create a skill', 'new agent', 'improve skill', 'skill creator', or 'agent creator'.

How to install AgentForge

Nothing to install: the source repository no longer has this skill. If the author brings it back, our daily re-check will pick it up and the command will reappear here.

Commands — how to trigger AgentForge

  • /agent-forge Guided authoring of OpenClaw skills and agents, with templates

It also activates on plain-language prompts like these:

  • Walk me through creating a new OpenClaw skill step by step
  • Help me build an agent with memory and self-improvement
  • Refine this existing skill using the guided 5-step process

Frequently asked questions

Is the AgentForge skill free?
Yes. The skill itself is free from AlekseiUL/agentforge-openclaw. SkillProof publishes the install command and an independent test verdict at no cost.
Does AgentForge work with Claude Code?
We tested it with Claude Code 2.x (agent harness) on Jul 24, 2026. Verdict: Works with setup. A Russian-language meta-skill for authoring OpenClaw skills and agents. Following the body produced a tighter skill scaffold than a freeform baseline: it enforces the workflow/role/data-driven/hybrid type taxonomy, puts triggers in the description, and its pitfalls table flags real traps like the unsupported 'version' frontmatter field and keeping data in data/ not memory/. The larger agent-creation half depends on the OpenClaw runtime (openclaw.json, gateway, sessions_send, cron) and could not be exercised or verified here. Clean on security: the body and templates contain no exfiltration, covert instructions, or self-escalation.
What is the AgentForge SkillProof Score?
7.6/10 — installs cleanly 5/5, triggers reliably 4/5, output vs. baseline 6/10, docs & honesty 4/5.
How do I install AgentForge?
Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
Can I use AgentForge with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.