automation-tool-router
Skill0 starsChanged 3 months agoMITView on GitHub
Choose the best automation layer across native connectors, Codex automations, MCP tools, Make.com scenarios, n8n workflows, and Starlight swarm queues. Use when a user asks which tool to use, how to automate a workflow, whether Make or n8n is better, how MCP should fit, how future agents should remember tool choices, or how to route recurring work safely.
In this repository8 other files it gives its agents.
- AGENTS.mdAGENTS.md
- codex-automation-operatorSkill
- make-com-operationsSkill
- mcp-governanceSkill
- n8n-operationsSkill
- starlight-excellenceSkill
- starlight-queen-queueSkill
- xquik-public-x-dataSkill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 8 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Distinctive: line 8
Use this first when the work is about automation architecture or tool choice.
- Distinctive: line 13
Use the native connector when one exists for the target system.
- Distinctive: line 14
Use Codex automation for recurring reasoning, review, repo, inbox, or briefing work.
- Distinctive: line 15
Use MCP when an agent needs a typed, bounded tool or resource.
- Distinctive: line 16
Use Make.com for fast SaaS glue, forms, CRM updates, and simple approval chains.
- Distinctive: line 17
Use n8n for branchy technical workflows, private data flows, reusable templates, and self-hostable automation.
- Distinctive: line 18
Use a Starlight queue for cross-repo, multi-agent, eval-heavy, or handoff-heavy work.
- Distinctive: line 25
Never print or commit secrets, live webhook URLs, customer data, inbox content, or private memory.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Automation Tool Router
Use this first when the work is about automation architecture or tool choice. Return a concrete owner, not a vague stack.
Routing Order
- Use the native connector when one exists for the target system.
- Use Codex automation for recurring reasoning, review, repo, inbox, or briefing work.
- Use MCP when an agent needs a typed, bounded tool or resource.
- Use Make.com for fast SaaS glue, forms, CRM updates, and simple approval chains.
- Use n8n for branchy technical workflows, private data flows, reusable templates, and self-hostable automation.
- Use a Starlight queue for cross-repo, multi-agent, eval-heavy, or handoff-heavy work.
Safety Gates
- Require human approval before send, post, spend, delete, publish, invite, production deploy, or account-permission changes.
- Dry-run first for migrations, workflow activation, bulk edits, and storage cleanup.
- Fail closed on auth, schema, transport, webhook, or confidence errors.
- Never print or commit secrets, live webhook URLs, customer data, inbox content, or private memory.
- Public exports must use mock payloads and pass a sensitive-marker scan.
Decision Output
Return:
- tool owner
- why this tool wins
- fallback tool
- trigger/input
- output contract
- approval gate
- validation step
- where durable state belongs
Reference Routing
- For Make.com scenario design, use
make-com-operations. - For n8n workflow design or exported workflow review, use
n8n-operations. - For MCP scopes, tool schemas, and server boundaries, use
mcp-governance. - For recurring scheduled work in Codex, use
codex-automation-operator. - For multi-agent queues and handoffs, use
starlight-queen-queue.
Verification
For public artifacts, run scripts/scan_public_safety.py from the source repo
when available. For installable bundles, run scripts/validate_skill_bundle.py.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

