agentleFS
link7373link7373/agentic-pmo/.claude/skills/design-dashboard/SKILL.md

design-dashboard

Skill1 starsChanged 3 months agoMITView on GitHub

Specify a portfolio dashboard — the decisions each audience tier must make, the semantic model and grain, a measure catalog with one definition per number, page layout, drill paths, refresh cadence and lineage. Produces a buildable spec, then hands off to /powerbi to build it when the platform is Power BI. Dispatches the portfolio-analyst with comms-lead.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When a portfolio reporting surface is needed or is being rebuilt, when two reports disagree about the same number, or when leadership keeps …

    line 9
  2. Distinctive:

    Write the measure catalog — every measure with a plain-English definition, formula, format and owner.

    line 31
  3. Very specific:

    Hand off to the build: If the platform is Power BI, run /powerbi — it builds the real project from this spec and powerbi-validator gates it.

    line 46
  4. Very specific:

    Start from templates/dashboard-spec.md.

    line 56

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

/design-dashboard — Specify the reporting surface

When to use

When a portfolio reporting surface is needed or is being rebuilt, when two reports disagree about the same number, or when leadership keeps asking for detail the current surface can't reach.

Dispatches

portfolio-analyst (lead) + comms-lead (the leadership tier's narrative and framing).

Inputs

knowledge/portfolio.md (what data actually exists and at what confidence), existing specs in knowledge/portfolio/, knowledge/product-context.md (goals to report against), knowledge/stakeholder-map.md (who consumes which tier), knowledge/integrations.md (platform and sync reality).

Steps

  1. Start from the decisions, not the data. List what each audience must decide, how often, and what they need to see to decide it. Name one to three questions the surface answers; more than three means it answers none of them well — split it. Every visual later earns its place by serving one.
  2. Pick the tier — operational (managers see their own data and gaps), tactical (themes, collisions, capacity), or executive (exceptions, decisions, investment vs. strategy). One semantic model, different altitudes — never a second dataset for a second audience.
  3. Define the grain explicitly (what one row means), then model it as a star schema with exactly one date table honouring the fiscal calendar in knowledge/cadence.md. Relationships single-direction unless bidirectional is justified in writing — ambiguous filter paths produce believable wrong numbers.
  4. Write the measure catalog — every measure with a plain-English definition, formula, format and owner. Resolve any term with more than one plausible reading ("active", "on track", "complete") to a single definition here; this table is the arbiter when surfaces disagree, and the name is a contract the builder implements character for character.
  5. Lay out pages and visuals against the questions from step 1, define the drill paths so every aggregate reaches its source rows in one or two steps, and specify the empty state. Give every page and visual a safe name — letters, digits, underscores, hyphens only. Leave chart formatting to the build target's design standard; specify what each visual must show and why, not how it looks.
  6. Specify refresh cadence, lineage, access, publishing rights, and failure behaviour — stale data must announce itself. Show data confidence on the surface, not just values.
  7. Work the privacy section deliberately. A portfolio surface carries sponsor and manager names, saved filter states can persist into the built report's files, and this surface reports the health of work — not the performance of people.
  8. Fill in ## Handoff: the build route, the capability tier actually available, and any open question blocking the build. Check knowledge/integrations.md for the platform first.
  9. Hand off to the build. If the platform is Power BI, run /powerbi — it builds the real project from this spec and powerbi-validator gates it. For any other platform the spec is the deliverable and a person builds it; say so plainly rather than implying a surface exists.

Methods

knowledge/methods/portfolio-management.md (reporting layers, dashboard design principles, KPIs), knowledge/methods/metrics-and-experimentation.md (metric hygiene, vanity measures), standards/communication-standards.md (audience playbook, BLUF, RAG discipline).

Output

Start from templates/dashboard-spec.md. A spec saved to knowledge/portfolio/dashboard-<surface>.md — a living document, updated in place so there is one current definition per surface. Log any measure-definition decision to knowledge/decision-log.md, since it binds every other report. Follow standards/document-standards.md.

The spec is complete when someone could build the surface from it without asking a further question. It is not "done" when it renders — only when every displayed number reconciles against an independent query and the empty state has been tested. Track that as an open item until the builder confirms it.

New measures go into knowledge/portfolio-measures.md before the build, never invented during it. That file is the contract the build implements character for character.


Created by Colin Beck<br> LinkedIn: https://www.linkedin.com/in/beckcolin/<br> GitHub: https://github.com/link7373

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.