Institutional Knowledge Archivist

When a long-tenured leader leaves, the org rarely loses the what — project trackers and docs survive. It loses the why: the decision rationale, the rejected alternatives, the stakeholder history, and the unwritten rules that lived only in that person’s head.

This skill turns a departing senior leader’s Microsoft 365 signals — emails, Teams chats, meetings, and SharePoint/OneDrive contributions — into a structured, interlinked institutional-knowledge archive a successor (or an AI assistant) can read on day one and ground on for years.

It runs as a guided, multi-session sprint rather than a one-shot dump. You curate and answer clarifying questions; the agent does all the writing.

Before you start

• A connected Microsoft 365 account for the leader (or a delegate working on their behalf). The skill reads mail, Teams chats, calendar/meetings, and recent files via the agent’s m365_* tools.
• Time: plan a ~1-week sprint across 5–6 sessions of 60–90 minutes. Session 1 does discovery; later sessions go deep, then assemble.
• Optional: a document skill installed if you want to export the final archive to .docx/.pdf.

Nothing leaves the M365 trust boundary — the archive is written to the user’s own Documents/ folder. HR, compensation, performance, and health content is automatically out of scope.

How to use it

Invoke it in plain language, e.g.:

• “Help me build an institutional knowledge archive for our departing VP.”
• “Capture Doug’s decision history before he retires next month.”
• “Start the institutional-knowledge skill.”

The skill then walks you through five gated phases, confirming with you before advancing each one:

1. Discovery — a complete inventory of the leader’s projects, initiatives, recurring forums, and workstreams across their whole tenure. You validate the significance ratings.
2. Deep extraction — per high-significance project: decision archaeology (what was decided, when, by whom, why, what was rejected, how reversible), problem→solution→lesson mapping, key relationships, and successor guidance.
3. Cross-cutting themes — leadership principles, org-navigation guidance, domain expertise, and historical context synthesized across projects.
4. Gap questionnaire — the questions only the leader can answer, generated from everything the signals couldn’t reveal. The leader fills these in offline.
5. Final assembly — a single, leader-approved archive document.

It’s resumable and idempotent: every session reads a manifest.md and picks up where the last one stopped, and re-ingesting the same source merges rather than duplicates.

Good to know

• The highest-value step is the clarifying-question loop. After each project deep-dive the agent asks you 2–3 targeted questions (“you and X exchanged 40 emails about Y but never wrote down the decision — what did you land on and why?”). That’s where tacit knowledge gets captured — don’t skip it.
• Breadth before polish. Early passes bias toward capturing raw signal; if you run out of time, raw breadth beats a half-polished narrow archive.
• Privacy is on you at share time. The archive includes private 1:1 and email signal by design (it’s the leader’s own data). The agent flags sensitive sections with [Review Required] / [Confidential], but the leader should review relationships.md, historical-context.md, and the “Successor guidance” sections line-by-line before sharing with anyone.
• Tool names may vary by host. The instructions reference m365_* tool names; if your Cowork/Scout environment exposes M365 under different names, the workflow is identical — just map the tool names.

What you end up with

institutional-knowledge/
├── phase-1-inventory.md   ← project & initiative catalog
├── projects/<slug>.md     ← one decision-rich deep-dive per project
├── decisions.md           ← cross-project decision archaeology
├── relationships.md       ← stakeholder roster + how to work with each
├── themes.md              ← leadership principles & operating style
├── domain-knowledge.md    ← technical/industry/regulatory knowledge
├── org-navigation.md      ← "how to get things done here"
├── historical-context.md  ← origin stories, past failures, unwritten rules
├── gap-questionnaire.md   ← what only the leader can answer
├── final-archive.md       ← the assembled, approved document
└── raw/                   ← faithful per-source extracts (audit trail)