Skip to content
Back to skills

Knowledge

ASecurity

Use when work surfaces something epistemic worth keeping — "we're assuming / proceeding as if X", "we decided X over Y", "open question worth outliving this session", "we're bound by X", "evidence for/against X", "testable hypothesis", or a stable concept worth naming — and when settling, relating, or surveying such records. Routes capture, settle, and survey intent to the correct `doctrine knowledge` verb.

  • 2 stars
  • 0 votes
  • 0 copies
  • 3 views
  • Added September 19, 2026
developmentgo

Works with

  • cli

Security analysis

A100/100

Scanned September 29, 2026

npx -y skills add davidlee/doctrine --skill knowledge --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Knowledge?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Knowledge
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/davidlee-knowledge/badge)](https://www.skillsdirectory.com/skills/davidlee-knowledge)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: knowledge
description: Use when work surfaces something epistemic worth keeping — "we're assuming / proceeding as if X", "we decided X over Y", "open question worth outliving this session", "we're bound by X", "evidence for/against X", "testable hypothesis", or a stable concept worth naming — and when settling, relating, or surveying such records. Routes capture, settle, and survey intent to the correct `doctrine knowledge` verb.
---

# Knowledge

Knowledge records are the **epistemic home** — typed, citable entities for
what the project believes, chose, wonders about, observes, or is bound by.
Records *shape* work; they are never work items themselves.

Four homes, one line: work intent → backlog; epistemic state → knowledge
record; architectural decision → ADR; agent guidance → memory. See
`lib:reference/using-doctrine.md` § Which home for which record.

The CLI is the source of truth for exact flags and each kind's status
vocabulary: `doctrine knowledge --help`.

## Which kind

| phrase shape | kind |
|---|---|
| "proceeding as if X" (unvalidated premise) | assumption (ASM) |
| "we chose X over Y" (scoped/operational) | decision (DEC) |
| "unresolved — X?" (needs an answer) | question (QUE) |
| "bound by X" (externally imposed limit) | constraint (CON) |
| "observed — X" (citable observation) | evidence (EVD) |
| "testable claim X" (awaiting evidence) | hypothesis (HYP) |
| "stable mental model / term X" | concept (CPT) |

**DEC or ADR?** Project-global with architectural consequences →
`doctrine adr new`; scoped or operational → DEC.

## Verbs

See `doctrine knowledge --help` (and `doctrine link --help` / `doctrine
supersede --help`) for the verb list and flags. Link a record to the work it
shapes with `shapes` (`spawns` for work it caused); an evidentiary edge is
`supports` or `disputes`.

## Wrong home?

- Work intent (bug / improvement / chore / risk / idea) → `/backlog`.
- Reusable agent guidance, recipe, or gotcha → `/record-memory`.

Discriminator: a knowledge record is a citable epistemic *entity* in the
relation graph — relatable, gateable, supersedable; a memory is
retrieval-layer guidance for agents. The same line splits CPT from a memory
`concept`: a CPT names a stable project concept other entities will cite; a
memory explains how to think or work.

## Gating — association is not gating

See `lib:reference/using-doctrine.md` § Which home for which record for the
knowledge-record gating mechanic (`needs`, settle-to-unblock).

## Rules

- See `doctrine knowledge show --help` — the id prefix resolves the kind on read
  and transition.
- Capture seeds the kind's default state (held, proposed, open, active, …).
- Don't hand-edit record TOML (`lib:reference/using-doctrine.md` § storage
  tiers) — use the verbs; the prose body (`*.md`) is hand-edited.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.