-21risk-automation

Overview of -21risk-automation skill

What -21risk-automation is for

The -21risk-automation skill helps an AI agent automate 21risk operations through Composio’s Rube MCP server. It is not a standalone script or local CLI workflow; it is an agent instruction layer for discovering the current _21risk tool schemas, checking the user’s 21risk connection, and executing the right Rube tool calls with less guesswork.

This is a good fit if you already use an MCP-capable client, need workflow automation around 21risk, and want the agent to avoid hard-coded assumptions about tool names or parameters.

Best-fit users and jobs

Use -21risk-automation for Workflow Automation when the job depends on live 21risk account access through Composio, such as creating, reading, updating, or coordinating 21risk tasks exposed by the _21risk toolkit. The skill is most useful for operators who want repeatable agent behavior: discover tools first, verify connection status, inspect schemas, then execute.

It is less useful if you only need a one-off natural-language explanation of 21risk, do not have Rube MCP connected, or cannot authorize a 21risk connection.

Main differentiator

The important differentiator is the “search tools first” pattern. Instead of assuming a fixed API schema, -21risk-automation instructs the agent to call RUBE_SEARCH_TOOLS before execution so it can retrieve current tool slugs, parameters, execution plans, and pitfalls. That matters because MCP tool schemas can change, and stale prompts are a common failure mode in automation skills.

How to Use -21risk-automation skill

-21risk-automation install context

Install the skill from the Composio skill collection in an environment that supports Claude-style skills and MCP tool access:

npx skills add ComposioHQ/awesome-claude-skills --skill -21risk-automation

Then configure Rube MCP in your client by adding the MCP server endpoint:

https://rube.app/mcp

The upstream skill itself lists no local scripts, helper files, or bundled references. The only source file to read first is SKILL.md at composio-skills/-21risk-automation/SKILL.md. For external toolkit behavior, check Composio’s 21risk toolkit documentation at composio.dev/toolkits/_21risk.

Required setup before usage

Before asking the agent to perform a 21risk action, verify three things:

1. RUBE_SEARCH_TOOLS is available in your MCP client.
2. RUBE_MANAGE_CONNECTIONS can manage the _21risk toolkit.
3. Your 21risk connection status is ACTIVE.

A practical first instruction is:

Use the -21risk-automation skill. First call RUBE_SEARCH_TOOLS for my specific 21risk task, then check _21risk connection status with RUBE_MANAGE_CONNECTIONS. Do not execute changes until you show me the discovered tool, required fields, and proposed action.

This reduces failed calls caused by missing authentication, wrong tool names, or incomplete inputs.

Turning a rough goal into a strong prompt

Weak prompt:

Update my 21risk records.

Better prompt:

Use -21risk-automation to update 21risk records for [business object or task]. First discover the current _21risk tools for this exact operation. Confirm the active connection, identify required fields, ask me for any missing IDs or values, then run the safest tool call. Summarize what changed and include any tool warnings returned by Rube.

The stronger prompt improves output because it tells the agent the target system, the desired operation, the discovery requirement, the approval boundary, and the expected final report.

Suggested operating workflow

For reliable -21risk-automation usage, follow this sequence:

1. Describe the exact 21risk job, not just the product name.
2. Require RUBE_SEARCH_TOOLS before any action.
3. Ask the agent to inspect the returned schema and list missing fields.
4. Use RUBE_MANAGE_CONNECTIONS to confirm _21risk is active.
5. Approve the final tool call only after the agent shows the intended operation.
6. Request a concise execution summary, including created or modified records if returned.

This skill is installation-light but context-sensitive: the quality of the result depends heavily on the specificity of your task and the live tool schema returned by Rube.

-21risk-automation skill FAQ

Is -21risk-automation beginner-friendly?

It is beginner-friendly for users who already have an MCP-capable client, but not for users unfamiliar with MCP connections. The skill explains the required Rube calls, yet you still need to authorize _21risk through RUBE_MANAGE_CONNECTIONS and understand when the agent is about to make a real account change.

How is this better than an ordinary prompt?

An ordinary prompt may tell the agent to “use 21risk,” but it may guess tool names, invent parameters, or skip connection checks. The -21risk-automation skill adds a concrete execution pattern: search current tools, confirm the _21risk connection, follow returned schemas, then execute. That makes it safer for live workflow automation than a generic instruction.

What can block adoption?

The main blockers are missing Rube MCP access, an inactive 21risk connection, or vague task inputs. Because the repository provides only SKILL.md and no scripts or examples for specific 21risk business processes, teams may need to create their own prompt templates for recurring operations.

When should I not use this skill?

Do not use -21risk-automation if you need offline processing, a fixed SDK integration, bulk automation with custom error handling, or a fully audited workflow without human review. It is best for agent-mediated tool execution, not as a replacement for a production integration service.

How to Improve -21risk-automation skill

Improve inputs before the first tool call

The biggest improvement is better task framing. Provide the target object, desired action, constraints, identifiers you already know, and whether the agent may write changes or must ask first. For example:

Find the current _21risk tool for [operation]. I have [known ID/value]. If a required field is missing, ask before executing. Treat this as a write operation requiring confirmation.

This gives -21risk-automation enough context to search for the right schema and avoid premature execution.

Watch for common failure modes

Common failures include skipping RUBE_SEARCH_TOOLS, using stale parameter names, assuming the connection is active, and treating an ambiguous user goal as permission to modify data. If the agent jumps straight to execution, stop it and restate the workflow: discover, check connection, validate fields, confirm, execute.

Iterate after the first output

After the first run, ask for a reusable prompt pattern based on the discovered tool schema:

Convert this successful -21risk-automation run into a reusable prompt template. Include required fields, optional fields, approval checkpoints, and the final summary format.

This turns a one-off automation into a repeatable operating procedure for your team.

Extend the skill for team use

If you maintain a local fork, improve -21risk-automation by adding examples for your most common 21risk workflows, safe-write approval rules, and field-mapping notes from real runs. Keep the core instruction to search tools first, but add organization-specific guardrails such as “never delete without explicit approval” or “always include the 21risk record ID in the final summary.”