WK
Wiki Knowledge Base Share Kit 10-skill knowledge-base + PM workflow package
GitHub repo Start Here Latest release
Structured knowledge, not note sprawl

Turn your
markdown vault into a
usable knowledge base.

A reusable open-source package with 10 installable skills and 9 coordinated workflow tracks. Start with knowledge-base maintenance first, then progressively load the PM mainline only when milestones, blockers, team alignment, or delivery gates become real needs.

Get the repository Read START-HERE Browse latest release
project knowledge ops task overview

The fixed page-role model stays stable. work-journal remains a separate workflow rather than a sixth page role.

Role-stable structure Keep project, knowledge, ops, task, and overview pages from drifting into each other.
Unified CLI install Clone the repo and run ./wiki-kit install. If you keep multiple runtimes on the same machine, run ./wiki-kit detect first and then choose --platform explicitly.
Default lane, optional PM lane Stay narrow for ingest, maintenance, and audit, then add project-management, team coordination, and delivery audit only when the work actually needs them.

How the package is organized

Think of the repository as a layered operating surface: onboarding first, knowledge loops next, PM layers when needed, collaboration beside them, and a stable vault shape as the output.

entry

Onboarding layer

Start from explanation-first guidance or low-friction setup.

knowledge-base-kit-guide Installation guide, profile setup, usage routing.
knowledge-base-orchestrator Environment check, vault skeleton, profile bootstrap.
→
core loops

Knowledge and PM specialist layer

These are the specialist loops that shape structure, durability, project progression, and auditability.

knowledge-base-ingest Long-form markdown split, linking, and baseline-driven import.
knowledge-base-maintenance Promote durable conclusions into the right page role.
knowledge-base-audit Metadata, links, orphan pages, noise drift, and governance checks.
knowledge-base-project-management Owner-side intake, milestones, blockers, priorities, and execution boards.
knowledge-base-delivery-audit Delivery completeness, evidence gaps, ready / blocked, and greenlight review.
→
collaboration

People and coordination layer

Support collaboration without turning the knowledge base into raw chat history or replacing the shared source of truth.

knowledge-base-working-profile Collaboration-relevant preferences, heuristics, and boundaries.
knowledge-base-team-coordination Questionnaires, alignment, assignments, and closeout decisions.
work-journal Daily execution notes that can later feed durable knowledge.
→
outcome

Stable vault output

The final target is not more pages, but a vault whose roles and navigation remain legible over time.

project knowledge ops task overview

Why this package exists

The point is not to write more notes. The point is to stop a vault from decaying into a maze of folders, mixed page roles, broken navigation, and hard-to-reuse knowledge.

Typical drifting vault Target state with this package
Root pages become timeline dumps full of temporary execution history. Root pages return to navigation, stable boundaries, and topic entrypoints.
ops pages become chronological logs of what happened once. ops pages become reusable symptom / root cause / fix / boundary references.
Long-form sources are imported as one big blob and become difficult to revisit. Sources are split into overview / chapter / topic structures with lineage and retrieval surfaces.
Collaboration context sits in chat, meeting notes, and personal memory. Working profile, team coordination, and journal workflows create a stable, reviewable collaboration layer.

What’s inside

The repository keeps the package explicit: 10 public skills, one recommended install script, one fixed page-role model, and a small set of repeatable workflows rather than a black-box automation promise.

onboarding

knowledge-base-kit-guide

Explains installation, profile setup, and which skill to use next.

orchestration

knowledge-base-orchestrator

Provides low-friction onboarding: inspect the environment, create a skeleton, generate a profile, and route the next step.

ingest

knowledge-base-ingest

Imports long-form markdown with a baseline-first, test-driven, lightweight-harness workflow.

maintenance

knowledge-base-maintenance

Writes durable conclusions back into the right page role while filtering one-off process noise.

audit

knowledge-base-audit

Inspects metadata, dead links, orphan pages, boundary drift, and noise regression.

project management

knowledge-base-project-management

Builds owner-side summaries, milestones, blockers, and personal execution boards without mixing management surfaces into governance.

team coordination

knowledge-base-team-coordination

Coordinates questionnaires, alignment, assignments, and closeout decision distillation.

delivery audit

knowledge-base-delivery-audit

Checks delivery completeness, evidence gaps, write-back gaps, and ready / blocked / greenlight conditions.

working profile

knowledge-base-working-profile

Distills collaboration-relevant preferences, heuristics, and boundaries without turning them into a private dossier.

work journal

work-journal

Captures daily work, meeting notes, and temporary ideas that can later feed durable knowledge.

default path

Start with knowledge-base loops

Most users do not need every workflow on day one. The default route stays narrow and stabilizes the vault first.

  • Kit Guide or Orchestrator for onboarding
  • Ingest, maintenance, and audit as the default loops
  • Working profile and journal when collaboration context needs continuity
optional PM mainline

Load PM only when delivery needs it

Project-management is now part of the package, but it stays progressive. Enable it when milestones, blockers, dependencies, handoff, or delivery gates become real needs.

  • knowledge-base-project-management for owner-side planning
  • knowledge-base-team-coordination for 2+ person alignment
  • knowledge-base-delivery-audit for ready / blocked / greenlight review

One method, several high-density scenarios

This package is strongest in environments that are knowledge-dense, frequently updated, collaboration-heavy, and in need of traceability and auditability.

Medical / pathology / research

Guidelines, textbooks, papers, and case summaries need structure, lineage, and careful update boundaries.

  • Long-form source ingest
  • Stable knowledge and ops pages
  • Review-friendly collaboration context

Enterprise knowledge bases

PRDs, runbooks, delivery docs, and meeting conclusions need reliable navigation, governance, and an optional owner-side delivery layer when execution starts to sprawl.

  • Maintenance and audit loops first
  • Optional PM boards and delivery gates
  • Cleaner onboarding for teammates

Cross-team collaboration

When context is fragmented across chat and temporary docs, shared coordination and journal workflows reduce repeated re-explanation.

  • Draft / approved boundaries
  • Traceable meeting and decision context
  • Working-profile hygiene with consent boundaries

Best-fit use-case poster

If you want one scenario that shows why this package exists, use this one: one shared vault for books, papers, SOPs, questionnaires, and project context.

Best-fit use case

One shared vault for
papers, books, SOPs,
questionnaires, and
project context.

Keep the same markdown/wiki vault on NAS or a synced drive, let each teammate continue using their own agent, and make long-form knowledge, intake forms, and delivery context flow back into one governed structure.

OpenClaw Hermes agent Claude Code NAS / cloud sync
Sources papers · textbooks · books · SOPs
Sync surface NAS / shared drive / cloud disk
Individual entrypoints each person keeps their own agent and working context
Output an auditable wiki instead of chat fragments and drifting folders
Result: people can improve the same wiki in parallel while still using their own agents to help with questionnaires, literature digestion, book notes, and working documents.
close-read ingest maintenance loop audit gate collaboration hygiene
input layer Papers / literature

Research papers, guidelines, and review notes enter the vault as structured source material instead of attachment piles.

Books / long-form markdown

Textbooks, manuals, and long markdown sources are split into chapter/topic layers with lineage preserved.

Questionnaires / intake forms

Each teammate can use their own agent to help fill forms, organize answers, and push reusable conclusions back into the wiki.

shared vault
Governed markdown wiki

One synchronized vault holds knowledge pages, ops pages, project context, questionnaires, and durable summaries while the package keeps metadata, links, and page roles coherent.

ingest maintenance audit
agent surfaces OpenClaw

Good for structured execution, source handling, and batch operations around the same shared vault.

Hermes agent

Fits ongoing collaboration, working context maintenance, and coordination-heavy routines.

Claude Code

Useful for large document work, script-assisted cleanup, and knowledge-structure refinement.

working-profile

Keep collaboration preferences and boundaries explicit instead of repeatedly re-explaining them.

team-coordination

Route questionnaires, assignments, and alignment through a shared contract instead of private chat memory.

work-journal

Let agents leave structured work logs so validated paths and dead ends can actually become reusable experience.

How it flows

Get the repo locally, run the installer, stabilize the profile contract, then stay on the default loops until PM workflows are actually needed.

01

Get the repo

Clone the repository or download a release archive so the package contents are local.

02

Run wiki-kit install

Use the unified CLI for Codex, Claude Code, Kimi CLI, or compatible agents platforms. If more than one skills directory is present, choose --platform explicitly. Manual folder copy is only the fallback path.

03

Choose onboarding

Use Orchestrator for a low-friction start or Kit Guide for an explanation-first path.

04

Prepare profile

Create or refine vault-profile.md. Add the project-management area only if you are actually enabling the PM mainline.

05

Run default loops first

Start with ingest, maintenance, and audit. Only add project-management, team-coordination, or delivery-audit when milestones, blockers, alignment, or delivery gates are truly part of the work.

fixed model

5 durable page roles

  • project for navigation, overview, and boundaries
  • knowledge for reusable methods and concepts
  • ops for troubleshooting and procedures
  • task for assignments and in-progress work
  • overview for indexes and governance surfaces
boundary note

Journal is not a sixth role

Journal content can be more fluid, but durable knowledge should be promoted into the fixed page-role model instead of remaining in the journal forever.

Get started from the right surface

Use the HTML page for overview, the repository for the actual package.

This landing page is designed for quick understanding. The GitHub repository remains the authoritative surface for install scripts, routing, templates, documentation, and release artifacts.

Open GitHub repository Go to START-HERE