Skip to main content
For AI agents: a documentation index is available at https://docs.coverbase.com/llms.txt. This page is also available in markdown by appending .md to the URL.
Alongside its tools, the server publishes prompts: scripted multi-turn workflows you invoke by name. Each one hands the assistant a sequence to follow, so you get the same shape of output every time instead of whatever the model improvises. How you invoke one depends on your client. In Claude, prompts appear in the attachment or slash menu once the connector is added. In other clients, look for “prompts” or “templates” in the MCP integration.
Prompts are a convenience, not a gate. Everything below is reachable by just describing what you want. The value is repeatability: the same review runs the same way each week, and the steps that are easy to skip don’t get skipped.
Prompts that lead to a write always instruct the assistant to get your explicit confirmation first. A prompt never grants standing permission to mutate data.

​
Vendor intake

submit_intake_request

Argument: vendor_nameFiles a new vendor intake request end to end: opens the session, reads your portal configuration, researches and confirms the vendor’s identity, surfaces existing vendors that already cover the need, runs your configured screening, and walks the inherent risk questionnaire one question at a time.See Vendor intake in chat for the full flow.
No arguments.Walks every pending-review intake session in turn: the requester’s use case and services, the enrichment and duplicate readout, the screening verdict, and the questionnaire answers.
Argument: intake_session_idApprove, reject, or request follow-up on one intake session.
Arguments: vendor_name, use_case (optional)Adds a bare vendor record with no due diligence. Only for when you explicitly want to skip intake. The assistant states what you’re giving up before it creates anything.

​
Assessments and controls

review_assessment

Argument: assessment_idWalks approving or sending back one assessment.
Argument: assessment_idWalks one assessment’s flagged issues, clears the false positives, and suggests guidance fixes so the same control stops misfiring next time.
Argument: days (default 180)Finds over-flagging controls across recent assessments and tunes their AI guidance to cut false positives. The companion to false_positive_review: one fixes the instance, this one fixes the cause.
Argument: assessment_idProduces an executive-ready report for a single assessment.
Argument: vendor_nameGenerates a pre-onboarding due diligence checklist for a new vendor.

​
Findings, evidence, and obligations

chase_overdue_findings

No arguments.Walks overdue and open findings, grouped so you can chase the right owners.
No arguments.Identifies and prioritizes vendors with expiring or missing evidence.
Argument: vendor_idAn evidence health audit for one vendor.
Argument: months_ahead (default 6)Contracts expiring in the next N months, with the vendor risk context attached.

​
Monitoring and investigation

triage_radar_alerts

Argument: days (default 7)Walks new radar alerts and what to do about each.
Argument: vendor_nameA detailed read of one vendor’s risk posture.
Arguments: object_type, object_idInvestigates the audit trail for a single platform object. The answer to “who changed this, and when”.

​
Reporting

weekly_briefing

Argument: days (default 7)A weekly TPRM executive briefing. The most-used prompt for a Monday standup.
No arguments.An executive risk briefing across the vendor portfolio.
Argument: days (default 90)A board-ready program report over a quarter-length window.

Example prompts

Free-form phrasing for the same workflows, when you don’t want to invoke a prompt by name.

Tool reference

The 11 tools these prompts drive, and every route each one covers.