Claude project prompt

Claude prompt for project risk register

Build or refresh a risk register from current project evidence, while keeping recorded risks separate from Claude’s inferred warning signs.

Best for: Weekly risk reviews, governance meetings, milestone readiness, and portfolio health checks.

Claude prompt

You are my project risk analyst. Use Kanera MCP to review the current project evidence and prepare a risk register.

Project or workspace: [NAME]
Review period: [DATE RANGE]
Important milestones: [MILESTONES OR "READ FROM KANERA"]
Scoring method: Likelihood 1 to 5 multiplied by impact 1 to 5
Risk threshold for escalation: [SCORE]
Time zone: [TIME ZONE]

Review active, blocked, overdue, and due-soon work. Inspect relevant comments, dates, dependencies, owners, and recent history. Look for schedule, scope, resource, quality, external dependency, and decision risks.

Create a risk register with these columns:
- Risk ID.
- Risk statement in cause, event, impact form.
- Evidence and Kanera card keys.
- Likelihood from 1 to 5, with a short rationale.
- Impact from 1 to 5, with a short rationale.
- Score.
- Trigger or early warning sign.
- Mitigation.
- Owner, only when recorded.
- Review date.
- Status: Recorded, inferred, accepted, mitigating, or closed.

Then provide:
1. The top three risks requiring attention.
2. Changes since the previous review when history supports them.
3. Risks that need owner or mitigation confirmation.
4. Suggested follow-up questions for the project manager.

Do not present an inferred risk as an established fact. Do not invent owners, costs, dates, or mitigations. Mark missing fields as "Needs confirmation." Keep the analysis read-only and cite source card keys for every risk.

Worked example output

Example result after Claude reviews a migration project board, due dates, blocked cards, comments, and recent activity.

Example result

Project risk register: Meridian migration
Review date: 18 September

R-01 | Data mapping approval delay
Risk statement: Because the legacy mapping has not been approved, the migration rehearsal may start with unresolved account rules, causing rework and schedule delay.
Evidence: MER-82 is overdue and blocked; MER-91 depends on its approval.
Likelihood: 4, approval is already three days late.
Impact: 5, the rehearsal and validation work depend on it.
Score: 20
Trigger: No approval by 19 September.
Mitigation: Escalate the decision and agree a fallback rule set for rehearsal.
Owner: Needs confirmation
Status: Recorded

R-02 | Test environment capacity
Risk statement: Because performance testing shares a constrained environment, the planned test window may be shortened, reducing confidence before launch.
Evidence: MER-104 is due in four days and its latest comment records a scheduling conflict.
Likelihood: 3
Impact: 4
Score: 12
Trigger: Environment booking not confirmed by 20 September.
Mitigation: Reserve a backup window and prioritise the critical scenarios.
Owner: Luca
Status: Inferred

R-03 | Unowned training content
Risk statement: Because the administrator training cards are unassigned, enablement material may miss the pilot start.
Evidence: MER-112 and MER-113 are active, unassigned, and have no activity in eight days.
Likelihood: 3
Impact: 3
Score: 9
Trigger: No owner assigned this week.
Mitigation: Confirm a content owner and reduce the first release to pilot-critical guidance.
Owner: Needs confirmation
Status: Inferred

Immediate attention
- Resolve R-01 before the rehearsal decision point.
- Confirm the environment booking for R-02.
- Assign ownership for R-03 and validate whether both training cards are required for pilot.

Get a more reliable result

Define your scoring scale and escalation threshold before asking Claude to rank risks.
Require a source card for every risk so the register remains auditable.
Treat inferred risks as review candidates until a project owner confirms them.

Prompt questions

Will inferred risks look like confirmed ones?

No. Each row is marked recorded, inferred, accepted, mitigating, or closed. Treat inferred risks as review candidates until a project owner confirms them.

Do I need a scoring scale first?

Yes. Define likelihood, impact, and the escalation threshold before asking Claude to rank risks.

Can this replace a RAID log?

It produces a risk register from live work. Assumptions, issues, and dependencies still need their own review if you keep a full RAID log.

Go from live Kanera work to a finished update.

Connect Claude to Kanera and refresh a risk register from live work instead of a stale spreadsheet or copied board snapshot.

Explore Kanera's AI workflow

Start free

30-day Pro trial. No card required.