All use cases

Work

Bot portfolio dashboard: every bot's role, skills, and routines on one Page

Once you run many bots, you lose track of who does what and which routines are on.

Share

What hurt

@keyelifeai asked Grok Bot for one living Page that shows each bot's name, purpose, skills, main routines, and status, and posted the full prompt. The prompt is strict about sources: a registry is the truth, skills are overlaid only from evidence, and nothing like health or last run is invented. Not affiliated with xAI.

What to install or create

One registry as source of truth → evidence-only skill overlay → wide filterable Page → refresh in place.

Bots

  • Chief of Staff

    Builds and refreshes the portfolio Page from the registry.

Plugins

  • None on day one. Follow the first-run steps.

Routines

  • None on the first run. Watch one manual pass first.

The prompt

Bot Portfolio Dashboard Page (author-quoted)

Quoted from @keyelifeai — as published. Do not invent a second system prompt.

Build me a living Bot Portfolio Dashboard as a Page I will return to and keep updated.

## What I want
A single operable view of my Grok Bot portfolio so I can see, at a glance:
- Bot name
- What it is for (role / durable outcome / scope)
- Which Skills it lands on (skill placements)
- Main Routines it runs (names + enabled/disabled)
- Operating status (status, lifecycle, operating mode)

This is an operations board, not a one-off org chart. Prefer a filterable wide Page over a long chat dump.

## Before you build
1. Restate in your own words what you think my goal and problem are. Wait for corrections if anything is off.
2. Find the Canonical / authoritative Bot Portfolio Registry (or the closest structured source of truth). Prefer that over free-text bot descriptions when they conflict.
3. Say clearly which fields come from the Registry and which must be overlaid from elsewhere.

## Source-of-truth rules (do not invent)
- Registry = source of truth for identity, role/scope, status/lifecycle/operating mode, primary routines, handoff/topology when present.
- Skills are usually NOT a Registry row field. Overlay Skill placements only from evidence (that Bot’s Getting started / named workflows / SKILL.md ownership). Label weak guesses or omit them. Shared Skills used by several Bots may list multiple owners with a “shared pattern” note.
- Do not invent: last successful run, acceptance, agent IDs, owners, handoff contracts, or “green” health when timestamps are null/stale.
- Empty Routines on an on-demand Bot are not automatically “inactive.” Zero Routines can be intentional—call that out if evidence says so.
- Distinguish freshness clocks if your Registry has them (e.g. last verified vs last live observation). Show as-of dates; never imply live verification you did not do.

## Dashboard shape
- Wide Page with filters (lifecycle / operating mode / status; default-hide retired/stub).
- Columns: Name · Function/scope · Skills (overlay) · Main Routines · Status · Routine on/off summary.
- Short caveat callout: what is Registry fact vs Skill overlay; what is unknown.
- Embed a dated snapshot JSON path/notes so future refreshes can diff SHA / last_verified_at instead of rebuilding from scratch.

## Keep it up to date
After the first Page ships, propose how to refresh it (owner Bot + cadence). Prefer editing the same Page in place—never spawn duplicate Pages for the same subject. Do not create a Routine that contradicts an existing governance HOLD (e.g. “this Bot keeps zero Routines”) without my explicit override.

## Delivery
- Chat: one or two lines + the Page link only (not the full table body).
- Report counts: operating / on_demand / retired / stub, Skill coverage gaps (Bots with 0 evidenced Skills), and any Registry vs live name mismatches.

First run

  1. 1Read https://x.com/keyelifeai/status/2113071324520448381 (credit @keyelifeai).
  2. 2Pick your main or Chief of Staff bot to own the Page.
  3. 3Paste the author-quoted prompt and correct its restated goal before it builds.
  4. 4Check which fields came from the registry and which were overlaid.
  5. 5Approve a refresh cadence that edits the same Page in place.

Approval fence

Read-only reporting; any new routine it proposes waits for your OK.

What to tweak

Steal the 'do not invent' rules: a dashboard that guesses is worse than no dashboard.

Source

keyelifeai @keyelifeai · Original post

Not affiliated with xAI. We did not write this prompt. Third-party template terms still apply.

Browse templates

Browse team blueprints

Pro pack

Want the full setup, not just the idea?

Pro packs give you the job text, prompts, schedules and pitfalls for popular use-cases.

See Pro