knowledge-doc
Keep a long recoverable document (a brainstorm, a decision log, an incident write-up) in `state/knowledge` of the private instance repository through `ummanu…
Skill by vladmesh
Chalk Labs
Need help deploying knowledge-doc?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/vladmesh-knowledge-doc
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: documents, writing
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md and code. Checked Oct 26, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/knowledge-doc
What's inside
- Skill name
knowledge-doc- Description
- Keep a long recoverable document (a brainstorm, a decision log, an incident write-up) in `state/knowledge` of the private instance repository through `ummanu knowledge write`, with no raw Git commands. Triggers are "write up this brainstorm", "keep a sprint log", "write up the incident", "save this reasoning", and any edit of a file under the instance repository's knowledge directory.
- License
- Apache-2.0
- Instructions
- 45 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Knowledge — the instance's long documents The instance repository's `state/knowledge/` is the place for long reasoning and context: brainstorms, decision logs, incident write-ups. The documents travel to the private remote with the rest of the checkpoint, so they survive a move to a new machine. ## What goes where - Executable work goes on the board, not here: sprints as entities (`open-sprint`), product issues through `ummanu issue`, and cards, which the sprint observer cuts. - The short current conclusion that must reach a head's context goes into curated memory (`state/memory/facts`), written by the curator. - The long reasoning and the context the conclusion came from goes here. Knowledge is not indexed, does not appear in `memory_search` and is never loaded into a head's context wholesale. A document is read on purpose when the history of a question is needed. ## How to write ```bash python3 -P -m ummanu knowledge write --instance <instance dir> --actor ummanu \ --path decisions/2026-07-25-sprint-1.md --file /tmp/sprint-1.md python3 -P -m ummanu knowledge delete --instance <instance dir> --actor ummanu \ --path runbooks/retired-runbook.md python3 -P -m ummanu knowledge list --instance <instance dir> ``` **Do not run raw `git add` or `git commit` in the instance repository.** The dispatcher's tick writer commits `state/board`, `state/runs` and `state/knowledge` into the same repository, and a manual commit races it. `knowledge write` takes the shared writer lock and writes only `state/knowledge`; the next tick carries the document out. - `--path` is relative, inside `state/knowledge`, and must end in `.md`. Choose the section by document type: `brainstorms/`, `decisions/`, `incidents/`. - `--file` is the full new text of the document: the command replaces the file wholesale rather than appending. When editing a document, keep a working copy (for example in `/tmp`), edit that and write it again. - The format is free; no frontmatter or metadata is required. - Writing the same content a second time returns `changed: false` and writes nothing. - A document whose subject is gone (a retired runbook, a dropped pointer) is removed with `knowledge delete` rather than marked outdated. The repository history keeps it. - A document containing a secret is rejected (a validation error, exit code 2) and never reaches disk. Remove the token from the text rather than working around the check.
From vladmesh/ummanu, Apache-2.0.
More from vladmesh/ummanu
- 0.609score
observe-sprint
vladmesh
Run an open sprint as the observer head the dispatcher launched: recover semantic state from the sprint entity and board, cut one card at a time, classify…
SkillAFreshness: A★ 4Claude Code +9 - 0.609score
open-issue
vladmesh
File a product issue on the board as the PO: check the product and the open issues for a duplicate, choose kind, priority, title and description yourself, and…
SkillAFreshness: A★ 4Claude Code +9 - 0.609score
retro
vladmesh
The retro agent's procedure — walk fresh head transcripts and the memory search log, find concrete failures (an answer given from a canon fact without a…
SkillAFreshness: A★ 4Claude Code +9 - 0.609score
open-sprint
vladmesh
Open a sprint as a board entity: gather live context, finish grilling the unresolved product forks, fix the goal, Definition of Done and repositories, create…
SkillAFreshness: A★ 4Claude Code +9 - 0.606score
curate
vladmesh
The memory curator's procedure — pull durable facts out of fresh transcripts, deduplicate them, and write them into the canon through `curator memory-write`.
SkillAFreshness: A★ 4Claude Code +9 - 0.331score
steward
vladmesh
No description from the publisher yet.
SkillAFreshness: A★ 4Claude Code +9
Signals
- Stars
- 4
- Forks
- 0
- Open issues
- 3
- Last commit (this folder)
- Oct 14, 2026 (16 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- Apache-2.0
- Repository
- github.com/vladmesh/ummanu
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from vladmesh
- 0.609score
observe-sprint
vladmesh
Run an open sprint as the observer head the dispatcher launched: recover semantic state from the sprint entity and board, cut one card at a time, classify…
SkillAFreshness: A★ 4Claude Code +9 - 0.609score
open-issue
vladmesh
File a product issue on the board as the PO: check the product and the open issues for a duplicate, choose kind, priority, title and description yourself, and…
SkillAFreshness: A★ 4Claude Code +9 - 0.609score
retro
vladmesh
The retro agent's procedure — walk fresh head transcripts and the memory search log, find concrete failures (an answer given from a canon fact without a…
SkillAFreshness: A★ 4Claude Code +9 - 0.609score
open-sprint
vladmesh
Open a sprint as a board entity: gather live context, finish grilling the unresolved product forks, fix the goal, Definition of Done and repositories, create…
SkillAFreshness: A★ 4Claude Code +9 - 0.606score
curate
vladmesh
The memory curator's procedure — pull durable facts out of fresh transcripts, deduplicate them, and write them into the canon through `curator memory-write`.
SkillAFreshness: A★ 4Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get knowledge-doc running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
knowledge-doc skill. Agent Mart. https://agentmart.dev/skills/vladmesh-knowledge-doc (last updated 2026-10-24)