portfolio-dashboard

name	portfolio-dashboard
description	Use when the user wants to generate, update, or customise an HTML dashboard from portfolio assessment data — produces a self-contained HTML file with level distribution, repo table, shared gaps, improvement plan, and trend visualisation from multiple quarterly assessments

Portfolio Dashboard

Generate a self-contained HTML dashboard from portfolio assessment data. The dashboard visualises AI literacy levels across repos with trend tracking from multiple quarterly assessments.

The output is a single HTML file with inline CSS and no external dependencies. It opens in any browser, works offline, and can be shared as an email attachment or Slack file.

Input

The skill reads portfolio assessment files from the assessments/ directory in the current repo. It looks for files matching *-portfolio-assessment.md.

If no portfolio assessment files exist, tell the user to run /portfolio-assess first.

Process

Step 1: Find Assessment Files

ls assessments/*-portfolio-assessment.md 2>/dev/null | sort

Report what was found:

• One file: Dashboard will show current state, no trends
• Two or more files: Dashboard will include trend visualisation
• No files: Stop and suggest running /portfolio-assess

Step 2: Parse Assessment Data

For each portfolio assessment file, extract:

• Date — from the filename (YYYY-MM-DD)
• Summary — repo count, assessed/estimated/not assessed counts
• Level distribution — count at each level (L0–L5), split by assessed vs estimated
• Portfolio median level
• Weakest discipline — name and average score
• Repo table — each repo with level, confidence, discipline scores, last assessed date
• Shared gaps — gaps affecting 3+ repos with recommended action
• Improvement plan — organisation-wide, cluster, individual items

For trend data (2+ files), also track per-repo level changes across assessments to show trajectories (e.g. L1 → L2 → L3).

Step 3: Generate the Dashboard

Produce a self-contained HTML file with these panels:

1. Summary header. Portfolio name (derived from the source description in the most recent assessment), date, total repo count, median level.

2. Level distribution. Colour-coded visualisation showing how many repos are at each level. Use the standard level colours:

Level	Colour	Hex
L0 — Awareness	Grey	#808080
L1 — Prompting	Light blue	#87CEEB
L2 — Verification	Blue	#4682B4
L3 — Habitat Engineering	Teal	#20B2AA
L4 — Specification Architecture	Green	#2E8B57
L5 — Sovereign Engineering	Gold	#DAA520

3. Trends (if 2+ assessments). Show how the portfolio has changed over time:

• Median level trend line
• Number of repos at L3+ over time
• Assessment coverage (% of repos with full assessments) over time
• Per-repo level trajectories in the repo table

4. Repo table. Each repo as a row with:

• Repo name
• Current level (colour-coded cell)
• Confidence (assessed / estimated / not assessed)
• Discipline scores (context, constraints, guardrails) — or "—" for estimated/not assessed
• Last assessed date
• Level trajectory (if multiple assessments show change)

5. Shared gaps. Table of gaps affecting 3+ repos with repo count and recommended action.

6. Improvement plan. Grouped by impact scope:

• Organisation-wide (50%+ of repos) — highest priority
• Cluster (2-4 repos)
• Individual

Step 4: Write the File

Save to assessments/portfolio-dashboard.html.

If a previous dashboard exists, overwrite it — the dashboard is regenerated from source data each time.

Step 5: Report

Dashboard generated: assessments/portfolio-dashboard.html

  Repos: N (N assessed, N estimated, N not assessed)
  Median level: LN
  Trend data: N quarterly assessments
  Shared gaps: N

Open with: open assessments/portfolio-dashboard.html

Design Constraints

• Self-contained: All CSS must be inline or in a <style> block. No external stylesheets, no CDN links, no JavaScript libraries.
• Readable without JavaScript: The dashboard must be fully readable with JavaScript disabled. JavaScript may be used for progressive enhancement (sorting, collapsing sections) but is not required for the content to be visible.
• Print-friendly: The dashboard should render well when printed or saved as PDF.
• No build step: The HTML file is generated directly. No npm, no bundler, no static site generator.

Customisation

After generating the initial dashboard, the user may ask for refinements. Common requests:

• Filter to a subset of repos
• Add threshold highlighting (e.g. repos below L2 in red)
• Add a dark mode toggle
• Make tables sortable
• Collapse the improvement plan into an expandable section
• Add a "last updated" timestamp footer

Apply these as modifications to the existing file. Keep the self-contained constraint — do not add external dependencies.