בס״ד

Reusable Prompt — Engagement Context Extraction for Claude / LLM Handoff

docs/HOW_TO/reusable-engagement-context-prompt.md · last changed (pre-VM history) · rendered from GitHub master

Reusable Prompt — Engagement Context Extraction for Claude / LLM Handoff

Codified from Gmail draft 19db6fe1f7b1abe7 · Apr 22 2026

This is the prompt template Sam pastes into a new Claude / Claude Code session
when he needs to produce a portable consulting-engagement context document
(the same format as the Eden Gardens EG-CONTEXT-PACK v3).

How to use this

  1. Start a new Claude or Claude Code session
  2. Paste everything below the divider line as the opening prompt
  3. Provide the inputs the prompt asks for (signed agreement, emails, artifacts, dates)
  4. Claude produces the portable context document

PASTE EVERYTHING BELOW THE LINE INTO ANY NEW CLAUDE CHAT OR CLAUDE CODE SESSION TO PRODUCE A PORTABLE CONTEXT DOCUMENT FOR A CONSULTING ENGAGEMENT IN THE SAME FORMAT AS THE EDEN GARDENS ONE.

===============================================================

You are being given the task of reconstructing the full context of a consulting engagement I'm involved in, for the purpose of producing a single portable Markdown document that any future LLM session can pick up and immediately act on.

INPUTS I AM PROVIDING (or will provide on request):

  1. The signed agreement — PDF, photo, or pasted text. If there is no signed document, I'll describe the verbal or implicit scope.
  2. Any email, memo, or written update I've sent the client during the engagement.
  3. Any working artifacts — action registers, command centers, org maps, trackers — in Drive, GitHub, or pasted inline.
  4. Access to my Gmail, Drive, and GitHub where relevant — search directly rather than asking me for things you can fetch.
  5. The date we're treating as "today" and any 30/60/90 (or equivalent) checkpoints that happened or didn't.

YOUR OUTPUT — a single Markdown document with these sections, IN THIS ORDER:

  1. Purpose of this document — one paragraph framing why the file exists (portable context for any LLM instance).
  2. Principals & Stakeholders — everyone involved, role, email/handle.
  3. The Engagement as Signed — title, counterpart signatures + dates, effective date, nominal end date, today's date + day number, verbatim purpose statement.
  4. Commercial Terms as Signed — fee, paid, outstanding, expense treatment, scope-beyond language, any open questions flagged in margins.
  5. The Signed Deliverables (+ Handwritten Annotations) — number each; typed content; explicit capture of handwritten marginalia and how they modify depth.
  6. Interpretive Framework — what the framing word means ("skeleton," "sprint," "advisory," etc.); what depth "deliverable" implies; structural relationships; the contractual success ceiling; scope-protection language.
  7. Key Communication Artifacts — memos, emails, deliverables sent to client; date, recipients, attachments (named + located), reply status.
  8. Self-Reported Progress Status — if a progress memo exists, summarize area by area.
  9. Current State as of [today] — for each signed deliverable: Delivered / In Progress / Deferred / Not Started; % complete; artifact location.
  10. Off-Scope / Off-Page Work Absorbed — what got done that wasn't in the signed document; for each, flag architectural (arguably in-scope by implication) vs. operational (genuinely out-of-scope).
  11. Dual-View — how the engagement looks from the client's seat AND from mine, with the less-sympathetic-to-me view given its strongest honest articulation.
  12. Neutral Third-Party Verdict — what a senior operator with no skin in the game would say to each side.
  13. The Move — the single strategic action that converts current ambiguity into a clean true-up.
  14. Artifact Inventory — table: item / location (Drive path, URL, or email date) / status (draft / final / sent / acknowledged).
  15. Open Holds — two lists: blocked on them (with dates first flagged) and blocked on me.
  16. Open Strategic Question — the one genuinely-open question I still need to decide or surface.

STYLE CONSTRAINTS:

WHEN YOU FINISH:

  1. Save the document as a Markdown file in /mnt/user-data/outputs and present it for download.
  2. Draft an email to my own address (the one visible in my Gmail) with the full document in the body and a searchable subject like: "[Engagement Name] — Portable Context for LLM/Claude Code Handoff (Day N)".

WHY THIS EXISTS:

I run engagements where the signed scope doesn't fully capture what I end up doing. I need a way to drop any future AI session into full context so I can ask strategic questions, draft follow-ups, or prepare true-up conversations without re-explaining the whole history. This prompt is my on-ramp.

SPECIAL NOTE ON ENGAGEMENTS WITHOUT A SIGNED DOCUMENT:

If there is no countersigned SOW, substitute Section 3 with "The Engagement as Understood" — reconstruct the implicit scope from emails, messages, and verbal record, flagging every item's source. Section 4 becomes "Commercial Terms as Established in Practice" — what was actually paid, when, for what; gaps to resolve. Everything else stays the same. The absence of a signed doc is itself a strategic fact to surface in Section 13 (The Move).


When to use this template

Lineage

The Eden Gardens Context Pack v3 (Gmail draft 19db92caa0d60b8a) was the first artifact produced from this template. It served as the proof-of-concept for the format.

Source trail · docs/HOW_TO/reusable-engagement-context-prompt.md @ master · rendered 2026-07-02 7:23 PM EDT by scripts/build-docs.py · the .md in the repo is the truth; this page is the phone-readable view