Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/exponen-agi/recursive-agentic-improvementsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/exponen-agi/recursive-agentic-improvements/create-agent)<a href="https://agentmods.dev/commands/exponen-agi/recursive-agentic-improvements/create-agent"><img src="https://agentmods.dev/badge/commands/exponen-agi/recursive-agentic-improvements/create-agent/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/exponen-agi/recursive-agentic-improvements/create-agent"><img src="https://agentmods.dev/badge/commands/exponen-agi/recursive-agentic-improvements/create-agent.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.06197 |
| Opus 5.5 | $0.00000 | $0.02479 |
| Sonnet 5.5 | $0.00000 | $0.01239 |
| Haiku 4.5 | $0.00000 | $0.00620 |
Grade A, and why
create-agent scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 755 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/create-agent — Research, Plan, and Scaffold a New AI Agent
Build any AI agent for any domain in any supported framework. The skill researches what tools and patterns actually exist before writing a single line of code.
Usage: /create-agent [framework]
Examples:
/create-agent agno→ then describe "a travel assistant that books flights and hotels"/create-agent crewai→ then describe "a competitive intelligence crew"/create-agent langgraph→ then describe "a DevOps automation agent"/create-agent google-adk→ then describe "a medical appointment scheduler"/create-agent→ will ask everything
Supported frameworks: Agno · CrewAI · LangGraph · Google ADK
Step 1 — Gather Requirements
If a framework is given in $ARGUMENTS, use it. Otherwise ask.
Ask all remaining questions at once — do not ask one by one:
- Which framework? Agno / CrewAI / LangGraph / Google ADK
- What should this agent do? Describe freely — domain, job, users, and goal. No need to match a preset category. Examples: "a travel assistant that searches flights and hotels", "a legal document summariser", "a customer support bot for a SaaS product", "a multi-agent DevOps pipeline that monitors, diagnoses, and fixes CI failures".
- Who are the users? Internal team / end customers / developers / automated system
- What tools or external services do you know you need? Leave blank if unsure — the research phase will discover options.
- Memory across sessions? Yes / No
- Should it work standalone or as part of a multi-agent system? Standalone / Multi-agent
- Agent name and slug — human name and kebab-case slug (e.g.,
travel-assistant) - Generate Test Suite? Yes / No (Default: Yes, conforming to the Test Constitution at
tests/TEST_CONSTITUTION.md)
Step 2 — Validate Environment and Check MCP
Framework package check
Run the appropriate check:
# Agno
python -c "import agno; print('agno', agno.__version__)"
# CrewAI
python -c "import crewai; print('crewai', crewai.__version__)"
# LangGraph
python -c "import langgraph; print('langgraph', langgraph.__version__)"
# Google ADK
python -c "from google.adk.agents import LlmAgent; print('google-adk ok')"
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago Changed 2bc74d5ba7aa
- 1mo ago Changed · +16 lines 5eada4304f0b
- 1mo ago Changed · +5 lines ba7cabda3bee
- 2mo ago First seen · 734 lines · 0 tokens per session scan A 23aed430b39e
create-agent is a command published in the GitHub repository exponen-agi/recursive-agentic-improvements (2 stars, last pushed yesterday), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 6,197 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
create-agent compared with commands from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| create-agentexponen-agi/recursive-agentic-improvements | Command | 2 | yesterday | MIT | A | 0 | Claude Code |
| checklistgithub/spec-kit | Command | 148k | today | MIT | A | 13 | Claude Code |
| clarifygithub/spec-kit | Command | 148k | today | MIT | A | 27 | Claude Code |
| specifygithub/spec-kit | Command | 148k | today | MIT | A | 13 | Claude Code |
| convergegithub/spec-kit | Command | 148k | today | MIT | A | 35 | Claude Code |
| speckit.bug.testgithub/spec-kit | Command | 148k | today | MIT | A | 13 | Claude Code |
| analyzegithub/spec-kit | Command | 148k | today | MIT | A | 26 | Claude Code |