document-management-plan

name	document-management-plan
description	Family office: Produce a document management plan — DMS choice, folder taxonomy, retention rules, access control.

Family Office · Document Management Plan

For Claude: How to Use This Skill

Skill instructions are preloaded in context when this skill is active. Do not perform filesystem searches or tool-driven exploration to rediscover them; use the guidance below directly.

When to Use

Invoke when the advisor asks about:

• document management
• DMS plan
• folder taxonomy
• document retention

Customer Pain (VOC)

Synthesized from family-office operator interviews (see references/voc-evidence.md for full provenance):

• Cybersecurity and the underlying technology stack are now top-of-mind for document handling, not afterthoughts — [dLoQJfqGgag @ 13:56]

Note: VOC depth on this skill is thin — Track-A2 lexicon mining surfaced one direct hit from the family-office interview corpus. v2 will mine DMS RFP corpora and incident-response postmortems directly to fill the gap.

Artifact Specification

This skill produces a single named deliverable:

• Artifact: Document Management Plan
• Format at this iteration: markdown (artifact.md) plus a structured interview.json capturing every advisor input. PDF / DOCX / XLSX companion renders and DMS / Snowflake push land in the execution-pipeline PR tracked under issue #427.

The artifact is written to a skill-local path, not a global directory:

<invocation-cwd>/artifacts/family-office/document-management-plan/<YYYYMMDD-HHMMSS>/
  artifact.md
  interview.json
  manifest.json

Interview Inputs

Minimum-viable interview — the skill asks only what is needed to personalize the deliverable. Pre-fill rules against family memory land in the execution- pipeline PR.

• dms_platforms — DMS platforms in use (SharePoint / Egnyte / Box)?
• top_level_taxonomy — Top-level folder taxonomy?
• retention_defaults — Retention defaults by document kind?
• access_roles — Access roles (principal / COO / advisor / staff)?
• audit_cadence — Access audit cadence?

Workflow

1. Run python scripts/agent.py --config config.json (or invoke via Claude Code with a config blob).
2. The agent validates config, runs the interview (TTY or fixture-driven), and produces the artifact under the canonical local path.
3. The agent writes manifest.json with artifact metadata (name, version, content hash, pillar, skill_name, created_at).
4. The agent optionally writes memory entries to the knowledge skill's memory_objects table via psycopg if config.memory_dsn is provided. Without a DSN, memory writes are skipped cleanly.

Memory Conventions

Memories written by this skill are tagged with:

• subject = the artifact name
• source = "document-management-plan"
• memory_type ∈ {decision, assumption, commitment, open_question}

Security & Confidentiality

• Never log interview answers or artifact contents at INFO level.
• Never include SSN, EIN, account numbers, or full financial amounts in log lines. If a WHERE-clause field carries such data, log only a sha256 hash.
• The artifact directory is local-only at this iteration. DMS push (with confidentiality-label routing) is handled by the execution-pipeline PR.

Reference

• Design spec: 20260419_Family_Office_Skill_Claude_Desigh.md
• Implementation plan: 20260420_FamilyOffice_Skills_Plan.md
• Catalog rebuild tracking: https://github.com/serenorg/seren-skills/issues/427