Guide · Agent skill
How to use the KPI Reporting agent skill (kpi-reporting)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The kpi-reporting skill teaches an AI agent to turn raw business or product metrics into a status readout for leaders, such as a weekly, monthly or quarterly business review (WBR, MBR, QBR), a scorecard, or an executive summary. This guide explains what the skill makes the agent do, how to install it, how to trigger it with plain requests, and what it needs before it will produce a report. It is written for people who have not used agent skills before.
KPI ReportingXiaomiMiMo/MiMo-Code · MIT Download the skill Quick start
- 1Confirm your data access. Make sure the agent can reach a connected structured data source, such as a warehouse or semantic layer, that holds your actuals.
- 2Download the ZIP. Download the kpi-reporting skill as a ZIP file from the page where you found this guide.
- 3Install the skill. Unzip it into ~/.claude/skills/ for personal use or into .claude/skills/ in a project, or upload it under Settings → Capabilities → Skills in the Claude apps.
- 4Make a natural request. Describe the readout, such as a WBR, MBR, QBR, scorecard or executive summary, with the period and audience.
- 5Provide the target basis. Answer any clarifying questions about the target, pacing or comparison period, since the skill asks before treating a basis as authoritative.
- 6Review the output. Check the status labels, the comparison basis, the validated drivers and the caveats before you share the readout.
What the KPI Reporting skill makes the agent do
The skill is a set of instructions that an AI agent reads when a task matches its description. Its job is to report status: pull actual numbers, compare them against targets or pacing, check whether explanations for movement are validated, and state what the team should do next.
The skill tells the agent to reproduce the topline actual before explaining any movement, and to show the current value, the absolute and relative change versus the comparison period, and a short interpretation for each headline KPI. Where a deadline applies, it asks the agent to show whether the metric is on pace to hit its target by the end of the period, not only whether it is above or below target.
- Confirms the period being evaluated, the comparison or target, and the data freshness cutoff.
- Pulls actuals from connected structured data sources before drafting anything.
- Labels status as red, yellow or green when useful, and states the comparison basis behind each label.
- Explains drivers only when they are validated, and otherwise names the next validation step.
- Ends with an operating implication: whether the movement is concerning and whether the main KPI is on track, at risk or ahead of plan.
What you need before you install it
The skill depends on access to a real source of truth. It expects connected structured data sources, such as a data warehouse or semantic layer, so the agent can query live actuals. Company docs, team communication tools and BI dashboards are optional and are referenced in the instructions as placeholders that your own setup can fill.
The skill also names three other skills that it can hand work to: design-kpis, metric-diagnostics and analyze-data-quality. Those are referenced by name in the instructions, so you should expect the agent to look for them when a task needs them.
- Required: access to connected structured data sources for live actuals.
- Optional: company docs, team communication tools and BI dashboards.
- Named in the instructions: design-kpis, metric-diagnostics, analyze-data-quality.
- If the required source is unavailable, the skill tells the agent to stop and ask for access rather than use weaker substitute data.
Install the kpi-reporting skill
There are two routes. You can download the skill as a ZIP file from the page where you found this guide, then unzip it into one of the skill folders below. Use the personal folder to make the skill available to you across projects, or a project folder to limit it to one project.
If you use the Claude apps, you can instead upload the skill under Settings → Capabilities → Skills.
- Personal install: unzip the folder into ~/.claude/skills/.
- Project install: unzip the folder into .claude/skills/ inside your project.
- Claude apps: go to Settings → Capabilities → Skills and upload the skill.
~/.claude/skills/kpi-reporting/Check the files in the skill folder
The package contains three files. SKILL.md holds the main instructions and the workflow the agent follows. references/report-templates.md holds starting patterns for the final artifact, such as an inline written update, a document or a single slide. agents/openai.yaml is also included in the package.
The report templates are starting patterns, not mandatory layouts. They only shape the output after the analysis is complete.
- SKILL.md: the main instructions and workflow.
- references/report-templates.md: templates for inline updates, documents or reports, and single slides.
- agents/openai.yaml: included in the package.
kpi-reporting/
SKILL.md
references/report-templates.md
agents/openai.yamlHow to trigger the skill with natural requests
You do not need a special command. Describe the readout you want in plain language, and include the period, the audience and the metrics if you know them. The skill's description covers requests to report status, compare against targets, explain validated drivers and state operating implications.
If your request is missing something that would change the conclusion, such as the target basis or the comparison period, the skill tells the agent to ask you before making precise claims.
- Prepare this month's MBR for the growth team, with status against plan for activation and retention.
- Build a QBR scorecard for the product metrics, with red, yellow and green labels against target.
- Write a short executive summary of last week's revenue numbers and what we should do next.
- Update the WBR for the checkout funnel and tell me whether conversion is on pace for the quarter target.
What the output looks like
The skill shapes the output to the format you need. A short chat, email or memo update leads with the answer, then gives headline actuals, comparison or pacing context, main drivers and the next step. A document or report is routed through a separate report-building skill named in the instructions, with a summary, KPI snapshot, driver analysis, business context, implications, caveats and source or calculation notes.
The skill refuses to draft a WBR, MBR, scorecard or KPI update from placeholder numbers. If you explicitly ask for a template or mockup, it can produce one.
- Inline written update for chat, email or memo.
- Document or report, routed through the report-building workflow.
- Single slide, following the surface you request.
Safety notes: scripts and network access
The skill does not run scripts. It does use the network, because its purpose is to query connected data sources for live actuals. Only connect sources you are comfortable letting the agent read, and check that the access you grant matches what the readout needs.
The skill also tells the agent to label data gaps and to avoid presenting unproven explanations as fact. Review the caveats and source notes in any output before you share it with leaders.
- Runs scripts: no.
- Uses network: yes, to query connected structured data sources.
- Limit the agent's data access to the sources the readout needs.
- Check the caveats, source notes and calculation method before sharing.
Tips for better results
Give the agent the period, the comparison basis and the target up front. The skill is built to anchor each readout in those three things, so a clear brief reduces back-and-forth. If you have a metric tree or a defined pacing model, point the agent to it.
If a driver explanation matters for the decision, ask the agent to validate it rather than accept a plausible story. The skill routes fresh driver investigation to the metric-diagnostics skill.
- Name the period, comparison period and freshness cutoff.
- Share the target or pacing definition, or say where it lives.
- Ask for the primary KPI first and keep supporting metrics to the smallest useful set.
- Request validation of drivers before the readout presents them as causes.
What people use it for
- Preparing a recurring weekly, monthly or quarterly business review with headline actuals and status against plan.
- Producing a product or growth KPI update with targets and red, yellow and green labels.
- Writing an executive summary that states the operating implication of a metric movement.
- Checking whether a KPI is on pace to hit a quarter-end target before the period closes.
Frequently asked questions
Is the kpi-reporting skill free to use?
The skill is published under the MIT licence. This page does not state any price, so check the licence text in the package for the terms that apply to you.
What do I need to run the kpi-reporting skill?
You need access to connected structured data sources, such as a data warehouse or semantic layer, so the agent can pull live actuals. Company docs, team communication tools and BI dashboards are optional.
Where do I install the kpi-reporting skill?
Unzip the downloaded ZIP into ~/.claude/skills/ for personal use or into .claude/skills/ in a project. In the Claude apps, you can upload it under Settings → Capabilities → Skills.
Does the skill run scripts or access the network?
The skill does not run scripts. It does use the network to query connected data sources for live actuals, so limit the access you grant to what the readout needs.
Can the skill explain why a metric moved?
It can present driver context only when the drivers are validated. For fresh driver investigation, the instructions route the work to the separate metric-diagnostics skill.
Will it create a KPI report if I have no data access?
No. The skill tells the agent to stop and ask for the required source rather than build a report from placeholder numbers, unless you explicitly request a template or mockup.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use kpi-reporting if you prepare recurring WBR, MBR or QBR reviews, or KPI status updates with targets, and your team has a warehouse, semantic layer or BI setup the agent can query. Pick something else if you need to design a metric system from scratch, which the skill hands to design-kpis, or if you have no real data source, because the skill will stop rather than substitute weaker data.
Download the skill