How to use these
Copy a prompt into Claude (or your AI assistant of choice), fill in the bracketed parts, run it, and iterate. These give you a strong first draft — review before you act, and never paste secrets or regulated data into a public model (see our Private LLM decision tool if that’s a constraint). Keep what works in your own context library.
Incident postmortem
What it’s for: a blameless postmortem that produces real fixes, not finger-pointing.
Draft a blameless postmortem for this incident:
[Paste timeline / logs / what happened]
Structure: summary, impact (who/what/how long), timeline, root cause (5-whys), what went well, what didn't, and 3-5 concrete action items with owners. Focus on systems and process, not people.System documentation
What it’s for: documenting a system so someone new can understand it.
Based on the details below, write clear documentation for [system/app]:
[Paste architecture notes / config / how it works]
Cover: what it does, key components and how they connect, dependencies, how to run/deploy it, common failure modes, and who owns it. Write so a new engineer could get oriented in 15 minutes.Runbook generation
What it’s for: a step-by-step runbook for a recurring operational task.
Write a runbook for [task, e.g., "restoring from backup" / "onboarding a new user"].
Include: prerequisites, step-by-step instructions (numbered, copy-pasteable commands where relevant), how to verify success, and what to do if a step fails. Assume the reader is competent but unfamiliar with this specific process.Ticket triage & summarization
What it’s for: turning a pile of tickets into a clear picture.
Here are [N] support/IT tickets:
[Paste tickets]
Group them by theme, flag anything urgent or security-related, identify the top 3 recurring issues (with likely root cause), and suggest one systemic fix that would reduce ticket volume most.Vendor / tool evaluation
What it’s for: a disciplined comparison of tools or vendors.
We're evaluating [tools/vendors] for [need]. Build a comparison on: fit to our requirements, total cost over 3 years, security/compliance posture, integration with our stack ([list]), switching/lock-in risk, and support quality.
Recommend one, with the reasoning and the main risk of being wrong. Practical for a mid-market IT budget, not enterprise.Access review
What it’s for: a periodic least-privilege check.
Here's our current access list for [system]:
[Paste roles/permissions/users]
Flag: anyone with more access than their role needs, stale accounts, shared credentials, and missing separation-of-duties. Output a prioritized remediation list (highest risk first).Tech-debt assessment
What it’s for: making the case for cleanup in business terms.
Here's a description of our current systems and known pain points:
[Paste]
Identify the top 5 sources of technical debt. For each: the business risk it creates (downtime, slow delivery, security), a rough effort estimate, and the cost of leaving it. Prioritize by business impact, and frame it so a non-technical owner understands the trade-off.Want this running across your environment?
This is a starter set. Inside an AI Office engagement, your team turns the highest-value of these into running workflows — documentation, triage, and monitoring inside your systems — built securely (we’re SOC 2 Type 2, with a Private LLM option when data can’t leave your walls). When one is worth shipping as a tool, a Value Sprint delivers it in weeks. Start a conversation at frogslayer.com.