agentleFS
pradeep-sankaran-07pradeep-sankaran-07/team-os-gdrive-template/.agents/skills/customer-dashboard/SKILL.md

customer-dashboard

Skill0 starsChanged 2 months agoMITView on GitHub

Build or refresh the shared customer dashboard: one self-contained HTML file covering every onboarding and live customer, with health, order or usage trends, blocked work and a per-owner summary. Creates it if it does not exist, updates it if it does. Triggers: "refresh the customer dashboard", "which clients are off track", "what are the red flags this week", "which accounts should I focus on", "I have a new revenue export", "update the client health dashboard". NOT for a single account's status (use account-review). NOT for a customer-facing report. NOT for writing up a call.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read the file named in revenue_meta.

    line 45
  2. Very specific:

    Every data drop keeps a .meta.json sidecar next to it recording source_file, exported_at, exported_by, normalised_at_utc, row counts, the …

    line 51
  3. Very specific:

    Write a dated snapshot to snapshots, and prune per config.yaml customer_dashboard.snapshot_retention.

    line 88
  4. Distinctive:

    No company-wide total, ever.

    line 107
  5. Distinctive:

    Never show a derived risk band on stale data.

    line 108
  6. Distinctive:

    Never delete an excluded row.

    line 109
  7. Distinctive:

    Never widen the Apps Script audience to the whole organisation.

    line 111

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

Customer dashboard

The only place in this OS where per-customer values sit side by side. That is a deliberate, named exception to the confidentiality rule, and it comes with two conditions that are not negotiable:

  1. Per-customer figures only.
  2. Never a company-wide total. A sum across all customers is company revenue, which stays out of every file including this one. Report risk as a share of the portfolio, never as an amount.

The exception covers portfolio/ and no other path.

Every path in this skill comes from config.yaml customer_dashboard. Read that block first and use output, crosswalk, revenue_csv, revenue_meta, snapshots and raw_drops as written there. Do not hardcode a folder name: the installer invites companies to rename these folders to their own vocabulary, and a hardcoded path silently writes to the old one.

Step 1 - Check the data, before using it

Read the file named in revenue_meta. If the export is older than config.yaml customer_dashboard.revenue_stale_days, suppress every derived risk band rather than showing a stale one, and say on the page that the data is stale and how old it is. A dashboard that silently shows last quarter's numbers as this quarter's is worse than one that shows nothing.

Every data drop keeps a .meta.json sidecar next to it recording source_file, exported_at, exported_by, normalised_at_utc, row counts, the period covered, rows dropped, and any assumption that is not confirmed (currency_assumed, currency_confirmed: false). Write the uncertainty down; do not resolve it silently.

Step 2 - Reconcile entities

The crosswalk file maps one customer across every system: account folder slug, display name, owner, tracker projects, roster status, renewal date, tier.

Rows that are deliberately out of scope are kept, prefixed _excluded_, with a written exclude_reason. Nothing is ever silently dropped, because a customer that vanishes from the dashboard looks identical to one that was never there.

Where two systems disagree, record the disagreement in a conflicts list rather than picking a winner in code.

Step 3 - Build

One self-contained HTML file. Inline CSS and JS, no external calls, so it opens from a Drive folder and from the Apps Script viewer alike.

Escape every value that came from account content. Account summaries are condensed from ingested email by a model, so their text is attacker-influenced: a customer name or a note containing a <script> tag would otherwise execute in the browser of everyone who opens the dashboard. Escape &, <, > and " in every interpolated value, put the data in a JSON island rather than building HTML strings, and add a restrictive Content-Security-Policy meta tag. Numbers from the value export are escaped too; they come from a spreadsheet a human edited. Apply acme-brand tokens for colour and type; charts use the brand palette, never a rainbow.

Sections: a portfolio summary, accounts off track, trends, blocked work, per-owner view, and a data health tab showing every source's age and every unresolved assumption.

Write a dated snapshot to snapshots, and prune per config.yaml customer_dashboard.snapshot_retention. Bump scoring_version whenever the scoring changes, so old snapshots are not read as comparable.

Step 4 - Share it

Google Drive's web preview does not run JavaScript, so a plain Drive link to this file renders blank. apps-script/serve-dashboard.gs is a ten-line Web App that serves it at a stable internal URL. Content changes need no redeploy.

Step 5 - Before anyone else can open it

The Apps Script viewer bypasses the Drive permissions: whoever can open the URL sees this file with the deployer's access. Deploy it to the named group in config.yaml customer_dashboard.apps_script.audience_group, never to the whole organisation, and have someone outside the group confirm they are blocked.

Non-negotiables

  • No company-wide total, ever.
  • Never show a derived risk band on stale data. Suppress it and say why.
  • Never delete an excluded row; keep it with a reason.
  • Escape every account-derived value before it reaches the page.
  • Never widen the Apps Script audience to the whole organisation.

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.