First Prompts by Role

Find your role, copy a prompt, and get a real result from your own cloud environment in about 10 minutes.

By the end of this tutorial you will have run one role-specific prompt against your live environment and refined it with follow-ups. Allow ~10 minutes.

What you’ll need: a workspace with at least one connection — CloudThinker can’t return real results without one. If you haven’t set that up, complete the quickstart first.

The syntax

Talk to agents in plain language. The full pattern is:

@agent #connection your request

• @agent (optional) — leave it out and CloudThinker answers, or mention one of your workspace’s custom agents.
• #connection (optional) — the connection to use, such as your AWS account or a Kubernetes cluster.
• / (optional) — run a skill or a command.
• Your request — what you need, in your own words. Add any context inline.

See Commands for the full list.

1. Open a chat

   Sign in at app.cloudthinker.io and click New chat in the left sidebar.

   Success state: an empty conversation opens. Type your request in plain language — no mention needed, CloudThinker answers.

2. Run your role's quick win

   Pick the tab that matches your role, paste the quick win, then try the follow-ups once you see the shape of the output.

   SRE / On-call

   Goal: Cut the time from alert to root cause. Get a structured investigation, not raw logs.

   You need: Pulse configured, plus connections for the systems you operate.

   Quick win — investigate an active alert:

   Investigate the current incident: pull related metrics, recent deploys, and topology, then propose the top 3 likely causes ranked by evidence
   

   CloudThinker splits the work across subagents, gathers evidence in parallel, and returns a ranked list of likely causes with the evidence behind each.

   Follow up:

   Draft a postmortem from this conversation
   Pull the matching runbook and walk me through the approval steps
   Check whether this pattern matches any past incident in memory
   

   DevOps / Platform

   Goal: Stop paying for cloud you don’t use. See where waste lives in under a minute.

   You need: an AWS, Azure, or GCP connection.

   Quick win — find idle and oversized resources:

   Find EC2 instances with <20% CPU utilization over the last 30 days, plus any unattached EBS volumes and unused Elastic IPs
   

   CloudThinker queries the cloud APIs, joins them with utilization metrics, and returns a ranked list with projected monthly savings.

   Follow up:

   Right-sizing for the top 5 by waste
   Cost trend by service for this quarter
   Draft a reserved-instance plan for the stable workloads above
   

   Security

   Goal: Find the highest-risk misconfigurations before an auditor or attacker does.

   You need: an AWS, Azure, or GCP connection.

   Quick win — surface public exposure on sensitive ports:

   List security groups with 0.0.0.0/0 access on database, SSH, or RDP ports across all regions
   

   CloudThinker returns the offending rules with resource owner, region, and a severity ranking.

   Follow up:

   SOC 2 compliance status with prioritized remediation
   Audit IAM policies for privilege-escalation paths
   Check for IMDSv1 instances or unencrypted EBS volumes
   

   Database

   Goal: Find the queries that are actually hurting you, not the ones you assume are slow.

   You need: a PostgreSQL or MySQL connection.

   Quick win — slowest queries in the last 24 hours:

   Show the top 10 queries by total time over the last 24 hours on production PostgreSQL, with execution count and P95 latency
   

   CloudThinker pulls from pg_stat_statements (or the equivalent), ranks by impact, and shows where the cost is concentrated.

   Follow up:

   Suggest indexes for the top 3 queries above
   Explain why query #2 isn't using the existing index
   Query latency P95 trends by endpoint
   

   Kubernetes

   Goal: See where pods are oversized, undersized, or imbalanced — before the next OOMKill or budget review.

   You need: a Kubernetes connection.

   Quick win — pod resource waste across the cluster:

   Analyze pod resource usage vs requests across all namespaces and surface the largest over- and under-provisioned workloads
   

   CloudThinker joins requests and limits with actual usage and ranks the gaps by node-cost impact.

   Follow up:

   HPA policies for the variable workloads above
   Find nodes with <30% utilization for consolidation
   Check for pods without resource limits or liveness probes
   

   Engineering leader

   Goal: Get one view across cost, security, performance, and reliability without scheduling four meetings.

   You need: connections in place for the areas you want covered.

   Quick win — quarterly infrastructure review:

   Run a quarterly infrastructure review covering top cost savings, security and compliance gaps, database performance hotspots, and Kubernetes utilization risk. Consolidate it into an executive summary with prioritized actions.
   

   CloudThinker works each area in parallel, deduplicates the findings, and returns a single executive brief instead of four tabs.

   Follow up:

   Turn this into a quarterly business review slide deck
   Track the open actions from last quarter — what shipped, what slipped
   Draft a roadmap that aligns cost reduction with reliability work
   

   Success state: CloudThinker replies naming resources, queries, or rules you recognize from your own environment — not generic best practices.

3. Refine the conversation

   CloudThinker keeps the conversation’s context, so build on the result instead of starting over:

   • Be specific upfront. “EC2 costs in us-east-1 for the last 30 days” beats “show me costs”.
   • Refine, don’t restart. “Drill into RDS for the items above” works.
   • Ask for the shape you want. Ask for a dashboard first to see the trend, then ask for recommendations on the same topic.
   • Ask across areas in one prompt. For anything that spans cost, security, databases, and Kubernetes, ask once — CloudThinker splits the work for you.

   Success state: a follow-up prompt returns an answer that references items from earlier in the conversation without you repeating them.

You’re done when…

• Your role’s quick win returned results naming resources or queries you recognize
• At least one follow-up prompt built on the first answer without restating it
• You can ask for a dashboard, a report, or recommendations in plain language

Next steps

Every prompt you just ran is one pass of the Detect → Analyze → Resolve → Validate loop described in How CloudThinker works — the modules below run that loop continuously. Set up your first module:

Resolve

Route alerts into Pulse so every incident is investigated before you open a terminal.

Review

Connect a repository and get review comments from @cloudthinker-ai on the next pull request.

Optimize

Run a scan to surface cost-saving findings and track them through implementation.

Skills

Teach CloudThinker your team’s procedures so it runs them your way.