Dossier logo

Dossier

Organization
pnp
dossier

Renders a polished, self-contained HTML briefing from any data source — SharePoint lists, uploaded documents, or a verbal description. Use when asked to build a competitor landscape, account brief, candidate summary, site inventory, board bio, speaker profile, partner overview, or any grid of profile cards or single-subject one-pager. Interviews the user if data is incomplete, constructs a validated JSON document, then renders it into a sandbox-safe HTML file using grid mode (2–8 subjects as responsive cards) or one-pager mode (1 subject as a rich briefing). No external dependencies — output runs inside a SharePoint sandboxed iframe.

Overview

Publisherpnp
Repositorysharepoint-skills
Skill namedossier
Stars
141
Forks
60
Bundled files
3
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by pnp on GitHub. Read the source before you install it.

Installation

Install the Dossier AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/pnp/sharepoint-skills/tree/main/Skills/dossier/dossier
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/pnp/sharepoint-skills.git /tmp/sharepoint-skills
mkdir -p .claude/skills
cp -r /tmp/sharepoint-skills/Skills/dossier/dossier .claude/skills/dossier
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Dossier in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Dossier on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Dossier is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Dossier

Generates a complete, self-contained HTML briefing by (1) interviewing the user or parsing their data, (2) constructing and validating a JSON document, then (3) rendering that JSON into HTML using assets/dossier-template.html. The JSON is the authoritative source of truth; the HTML is fully determined by it.


Step 1: Determine the Data Source

Check what the user has already provided before asking anything:

  • SharePoint lists: Call list_items on the relevant lists. Read all needed columns before proceeding.
  • Uploaded documents or tables pasted in chat: Parse all content. Do not ask for data you already have.
  • Verbal description only: Proceed to Step 2.

Step 2: Interview the User

If the data or intent is not fully clear, ask these questions in a single message — not one at a time:

  1. Subject type: What are you profiling? (competitors, accounts/prospects, candidates, SharePoint sites, board members, speakers, partners, vendors — or something else)
  2. Subject count: How many subjects? This determines the rendering mode.
  3. Purpose and audience: What decision does this briefing support? Who reads it? This determines detail level and which key facts matter most.
  4. Key facts: What are the 3–6 structured facts most important for each subject? Examples for competitors: HQ, founded, ARR, headcount. For candidates: current role, tenure, location, skills. For sites: owner, content type, last activity.
  5. Narrative emphasis: Should the summary emphasize product/capabilities, financials, relationships, risks, or something else?
  6. Palette: Warm paper (cream/terra-cotta), deep ink (dark/gold), clean white (white/green), or slate (gray/blue)? Default to warm paper.
  7. Mode override (only ask if count is 1 or user seems to want detail): Do you want a full one-pager briefing or a compact card?

Confirm your understanding before building the JSON.


Step 3: Choose the Rendering Mode

Auto-detection rule:

  • 2–8 subjects → grid mode
  • 1 subject → one-pager mode
  • User explicitly says "one-pager" or "full briefing" → one-pager mode regardless of count
  • User explicitly says "grid" or "landscape" → grid mode

Do not ask about mode unless the count is exactly 1 and the user has not indicated a preference.


Step 4: Decide What Goes Where

Key facts hold structured, scannable, low-word-count data — the kind of facts a reader needs in 5 seconds. Rules:

  • 3–6 pairs. No sentences. Values should be short (under 6 words).
  • Good: "HQ | Austin, TX", "ARR | ~$40M est.", "Founded | 2019"
  • Bad: "Company description | This is a company that..." (belongs in summary)

Tags are the fastest signal — 2–5 short labels that classify the subject. Use for: industry vertical, stage, geography, product category, or any grouping the reader uses for orientation.

Summary is 2–5 sentences of prose. Write in third person. Lead with the most important fact. No bullet points. This is the editorial paragraph — what an analyst would write.

Activity holds the 2–5 most recent, most relevant developments — in most-recent-first order. Each item gets a short date prefix (e.g., "Apr 2026") and a one-sentence description. For competitors: product launches, fundraises, acquisitions, executive changes. For candidates: recent projects, promotions, publications. For sites: content updates, ownership changes, traffic milestones.


Step 5: Build and Validate the JSON Document

Construct the JSON object. Top-level structure:

report object (document metadata):

FieldRequiredDescription
titleyesPage title, e.g. "Competitor Landscape — Q2 2026"
subtitlenoOptional subtitle shown below title
paletteyeswarm-paper, deep-ink, clean-white, or slate
modeno"grid", "one-pager", or "auto" (default: "auto")
footernoFooter text, e.g. "Prepared May 28, 2026 · Confidential"

subjects array — ordered list of subject objects (1–8 items):

FieldRequiredDescription
idyesUnique slug, e.g. "acme-corp"
nameyesSubject display name
typeyesCategory label shown as a pill, e.g. "Competitor", "Prospect", "Enterprise Site"
statusnoStatus label text, e.g. "Active", "Watch", "Inactive"
status_colornogreen, amber, red, gray, or blue
key_factsyesArray of {label, value} pairs, 2–6 items
tagsyesArray of strings, 1–6 tags
summaryyes2–5 sentence prose paragraph
activitynoArray of {date, text} items, 0–5, most-recent first

Validation Checklist

Run every check before rendering. Fix any failure before proceeding.

Document level:

  • report.palette is one of: warm-paper, deep-ink, clean-white, slate
  • subjects array has 1–8 entries
  • report.mode is "grid", "one-pager", or "auto" (or omitted)

Per subject:

  • id is present and unique
  • name and type are present and non-empty
  • key_facts has 2–6 items; each item has label and value
  • tags has 1–6 items
  • summary is present and 2–5 sentences
  • status_color (if present) is one of: green, amber, red, gray, blue
  • activity items (if present) each have date and text; at most 5 items

Data integrity:

  • No key_facts value is a sentence — values must be short labels
  • Activity items are ordered most-recent first
  • Subject count matches the chosen mode (grid: 2–8, one-pager: 1; or mode is explicitly set)

Step 6: Render the HTML

  1. Read assets/dossier-template.html.
  2. Replace {{REPORT_TITLE}} in <title> with a concise page title.
  3. Replace {{PALETTE_CLASS}} on <body> with palette- + the palette name.
  4. Determine final mode: if report.mode is "auto" or omitted, use "grid" for 2–8 subjects and "one-pager" for 1 subject.
  5. Replace {{PAGE_HEADER}} with the .rpt-header strip (title + optional subtitle).
  6. Replace {{CONTENT}} with the rendered subjects — either the grid of cards or the one-pager layout. See references/components.md for exact HTML patterns.
  7. Replace {{FOOTER}} with the footer text (or empty string).
  8. Deliver the complete file as the response. No prose outside the file.

Hard Constraints

  • No external URLs, CDN links, or web fonts in the output HTML.
  • No <script src> tags. All JavaScript must be inline.
  • All values must be pre-rendered — the page must be fully correct with JavaScript disabled.
  • Do not generate Python, PowerShell, or any other scripts — output is always the HTML file.
  • Do not render until the validation checklist passes.
  • Do not use SharePoint Person column types in any accompanying list schema — use text columns only.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Dossier AI skill do?

Renders a polished, self-contained HTML briefing from any data source — SharePoint lists, uploaded documents, or a verbal description. Use when asked to build a competitor landscape, account brief, candidate summary, site inventory, board bio, speaker profile, partner overview, or any grid of profile cards or single-subject one-pager. Interviews the user if data is incomplete, constructs a validated JSON document, then renders it into a sandbox-safe HTML file using grid mode (2–8 subjects as responsive cards) or one-pager mode (1 subject as a rich briefing). No external dependencies — outpu...

Why use Dossier on TypingMind?

Because you install it once and use it with any model. Dossier is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Dossier in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/pnp/sharepoint-skills/tree/main/Skills/dossier/dossier. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Dossier?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Dossier?

As many as you like. As long as a model supports skills, you can use Dossier with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Dossier AI skill free?

Yes. It is published on GitHub by pnp under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇