Lever Automation
by ComposioHQLever Automation helps AI agents work with Lever ATS through Composio MCP to list postings, browse opportunities, manage requisitions, stages, and tags with safe recruiting workflows.
This skill scores 72/100, which means it is acceptable for directory listing but should be presented as a focused integration guide rather than a fully self-contained automation package. Directory users get enough evidence to understand when to use it, how it connects through Composio/Rube MCP, and what kinds of Lever ATS actions it supports, though adoption still depends on external toolkit behavior and some setup guesswork.
- Clear purpose and trigger surface: it is specifically for automating Lever ATS recruiting workflows such as postings, opportunities, requisitions, pipeline stages, and tags.
- Operational guidance includes setup steps, required Composio/Rube MCP dependency, OAuth connection flow, and Lever API scope expectations.
- Core workflows expose concrete Lever tool names and parameters, such as `LEVER_LIST_POSTINGS` with filters for state, team, department, location, commitment, pagination, and tags.
- No repository support files, scripts, references, or README are present beyond SKILL.md, so users must rely on the single skill document and external Composio toolkit docs.
- Setup is described via MCP URL and OAuth prompt, but there is no explicit install command or verification step in the skill file.
Overview of Lever Automation skill
What Lever Automation does
Lever Automation is a recruiting operations skill for working with Lever ATS from an AI agent, especially Claude Code, through the Composio Lever integration. It helps users query and update hiring data such as job postings, candidate opportunities, requisitions, pipeline stages, and candidate tags without manually switching into the Lever UI for every action.
Best fit for recruiting and talent operations teams
The Lever Automation skill is most useful for recruiters, recruiting coordinators, talent operations teams, and hiring managers who already use Lever and need repeatable ATS actions. Good use cases include listing published roles, filtering openings by department or location, checking candidate pipeline status, moving opportunities between stages, and applying tags consistently across candidates.
What makes this skill different from a generic prompt
A normal prompt can draft recruiting text, but it cannot safely interact with Lever data unless the agent has the right tool access. This skill documents the Lever-specific tool context, expected operations, authentication flow, and practical parameters such as posting state, pagination, stage handling, and read/write permissions. That makes it better suited for real ATS automation than asking an AI model to “help with Lever” in the abstract.
Important adoption considerations
Lever Automation depends on the Composio MCP server via rube, so it is not a standalone script. You need a Lever account, OAuth/API access, and scopes that match your intended actions. Read-only reporting prompts require less permission than workflows that update opportunities, requisitions, stages, or tags. For production recruiting data, start with list/query workflows before allowing write operations.
How to Use Lever Automation skill
Lever Automation install and setup path
Install the skill in your agent environment with:
npx skills add ComposioHQ/awesome-claude-skills --skill "Lever Automation"
Then configure the Composio MCP server:
https://rube.app/mcp
When you run a Lever-related command, the agent should prompt you to connect your Lever account through OAuth. Confirm that your Lever API permissions include the resources you plan to access, such as postings, opportunities, requisitions, and write permissions if you want updates.
Inputs the skill needs for reliable results
For strong Lever Automation usage, provide the business goal plus the ATS fields that constrain the request. Useful inputs include posting state, team, department, location, commitment type, candidate name or opportunity ID, stage name, requisition context, tags to add or remove, and whether the action is read-only or should update Lever.
Weak prompt: “Show me engineering jobs.”
Stronger prompt: “Using Lever, list published engineering job postings in the San Francisco location, return title, posting ID, department, commitment type, and limit results to 50.”
The stronger version reduces ambiguity, avoids broad data pulls, and gives the agent a clear output shape.
Practical workflow for safe recruiting automation
Start with discovery, then act. First ask the agent to list or filter records, such as published postings or active opportunities. Next ask it to summarize the matching records and confirm IDs before making changes. Only then request updates such as moving a candidate to a pipeline stage or adding a candidate tag.
For write actions, include a confirmation step: “Preview the changes first and wait for approval before updating Lever.” This is especially important when modifying candidate opportunities, requisitions, or tags because small mistakes can affect recruiter workflows and reporting.
Repository files to read before using
The main file to inspect is SKILL.md under composio-skills/lever-automation. It contains the setup notes, supported workflow categories, and example tool usage such as listing job postings with LEVER_LIST_POSTINGS. There are no extra rules/, resources/, references/, or scripts/ folders in the current skill package, so adoption depends heavily on understanding the documented workflow and the Composio Lever toolkit docs at composio.dev/toolkits/lever.
Lever Automation skill FAQ
Is Lever Automation for Recruiting or general HR automation?
Lever Automation for Recruiting is focused on Lever ATS workflows: postings, opportunities, requisitions, pipeline stages, and tags. It is not a full HRIS automation layer for payroll, performance management, onboarding tasks, or employee records outside Lever.
Can beginners use the Lever Automation skill?
Yes, if they already understand their Lever workspace and have help from an admin for authentication and permissions. Beginners should start with read-only prompts such as listing postings or browsing candidate opportunities. Avoid write actions until you know which IDs, stages, and tags your Lever workspace uses.
When should I not use Lever Automation?
Do not use it when you lack permission to access candidate data, when your task requires legal or compliance review, or when bulk changes could affect active hiring pipelines without human approval. It is also a poor fit if your company does not use Lever or if your agent environment cannot connect to the Composio MCP server.
How does it compare with working directly in Lever?
The Lever UI is better for visual review, one-off recruiter decisions, and sensitive candidate evaluation. Lever Automation is better for repeatable queries, structured summaries, batch-like operational checks, and agent-assisted updates where the input criteria are clear. Many teams will use both: the skill for speed and consistency, the UI for final review.
How to Improve Lever Automation skill
Improve Lever Automation prompts with exact constraints
The fastest way to improve Lever Automation output is to specify scope precisely. Include filters such as published versus draft, department, location, hiring team, stage, tag, date range if supported by the tool, and result limit. Also state the desired format: table, CSV-style rows, grouped summary, or action plan.
Example: “Find active opportunities tagged Backend that are in the onsite stage, show candidate name, opportunity ID, current stage, owner, and next recommended follow-up. Do not update Lever.”
Avoid common failure modes in ATS actions
Common problems include using human-readable names when the tool needs IDs, requesting updates before confirming the matched candidate, pulling too many records without pagination, and assuming every Lever workspace uses the same stage or tag names. Ask the agent to list available stages or confirm matching records before modifying opportunities.
Iterate after the first output
Treat the first response as a narrowing step. If the agent returns too many postings or candidates, refine by department, location, owner, tag, or state. If it cannot identify a unique candidate, provide the opportunity ID or more context. If an update is proposed, ask for a dry-run summary showing old value, new value, target record, and reason before approving.
Add team-specific operating rules
For better long-term results, document your team’s recruiting conventions outside the skill: approved tags, stage transition rules, naming patterns for requisitions, and who can approve candidate updates. Then include those rules in prompts or local agent instructions. Lever Automation becomes more reliable when the agent has both tool access and your organization’s ATS operating policy.