Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/vscarpenter/gsd-task-manager/open-knowledgenpx skills add vscarpenter/gsd-task-manager --skill open-knowledgegit clone --depth 1 https://github.com/vscarpenter/gsd-task-managerWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00148 | $0.05179 |
| Opus 5 | $0.00074 | $0.02589 |
| Sonnet 5 | $0.00030 | $0.01036 |
| Haiku 4.5 | $0.00015 | $0.00518 |
Grade A, and why
open-knowledge scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 134 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OpenKnowledge — agent guidance
OpenKnowledge (OK) is a markdown-CRDT collaboration platform exposed via MCP. This skill is the single source of OK agent guidance. Every rule below is a MUST unless marked otherwise. Depth lives in references/*.md — one level deep; load a reference when its task comes up.
Skill version tracks
@inkeep/open-knowledge-server.cat ~/.ok/skill-state.ymlshows what's installed.ok seedneeds@inkeep/open-knowledge>= 0.4.0; if it errorsunknown command,npm install -g @inkeep/open-knowledge.
Setup (not connected yet?). If the
mcp__open-knowledge__*tools aren't available in your client, this project isn't wired up on this machine — seereferences/setup.mdfor the rung ladder (approve.mcp.json→ok startCLI → optional desktop app) and the canonical quickstart.
TL;DR — the 90% case
- Reads:
exec("cat …")for one doc,exec("ls -A …")for a directory (folder defaults + template menu),exec("grep …")for literal,searchfor ranked retrieval. NativeRead/Greponly on source code (.ts/.py/ …), never on in-scope.md/.mdx. - Writes:
write({ document: { path, content } })for a new or full-replace doc;edit({ document: { path, find, replace } })for a body find/replace;edit({ document: { path, frontmatter } })for a frontmatter merge-patch (nulldeletes a key).delete({ document })removes,move({ from, to })moves/renames. Body find/replace is body-only. Pass a one-linesummary(≤80 chars, user-facing outcome) on every content write. - Preview / open a doc — determine your ONE surface FIRST (once per session). Before opening anything, pick where to focus, stop at first match:
OK_DESKTOP_TERMINALset (check your env —echo "$OK_DESKTOP_TERMINAL") → OK Desktop's own terminal →ok open <name>(switches this window) · havepreview_*→ Codex Desktop pane · have a URL-navigation/browser tool → Cursor/Codex →preview_urlthen navigate it · else plain CLI →ok open <name>.ok open <name>opens a doc or folder (auto-detected — no--folder); add--skill <name>for a skill. ThepreviewUrlfield is a route id, not your open mechanism — don't hand it to a browser just because you saw it. Don'tpreview_screenshotto confirm edits. Full Step-0 procedure + per-surface how-to:references/preview.md. - Workflow guides:
workflow({ kind: 'ingest' | 'research' | 'consolidate' | 'discover' })returns a procedural guide, not data. Use it when the work fits the layer. - Direct questions: a plain business question ("which customers…", "what did we decide about…") routes to
search/exec+ a cited chat answer — no "research" keyword needed. Persist only when durable + multi-doc + not already covered, and offer first. Seereferences/corpus-qa.md. - Authoring or improving a skill ("write/make a skill", "improve this skill", "turn this into a skill"): STOP and invoke the
open-knowledge-write-skillskill — it owns scope choice (project vs global), the SKILL.md contract, evaluation, and install; don't improvise from this skill. Author through theskilltarget (write({ skill })), never a raw.ok/skills/…document. Never read, diff, or edit the installed projections under.Codex/skills/·.cursor/skills/·.codex/skills/·.opencode/skills/·.pi/skills/·.agents/skills/—installgenerates them and overwrites them on the next sync; they are NOT source of truth.
What ships with it
15 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/anti-patterns.md 6.7 KB
- references/cadence-and-logs.md 2.3 KB
- references/components-and-visuals.md 5.8 KB
- references/conflict-resolution.md 2.4 KB
- references/corpus-qa.md 2.4 KB
- references/doc-editing.md 3.3 KB
- references/folder-model.md 3.1 KB
- references/ingest-and-sources.md 907 B
- references/linking.md 4.0 KB
- references/media-and-assets.md 2.0 KB
- references/preview.md 11 KB
- references/setup.md 1.9 KB
- references/template-authoring.md 4.0 KB
- references/workflow-guides.md 3.8 KB
- references/writing.md 1.7 KB
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 134 lines · 148 tokens per session scan A 5f8839369573
open-knowledge is a skill published in the GitHub repository vscarpenter/gsd-task-manager (24 stars, last pushed 4d ago), licensed MIT. It adds 148 tokens to every session and 5,179 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
organize-database
Transform unstructured Notion pages into a well-designed database with proper schema and migration.
db-change
Use when writing SQL, adding a column or table, or touching anything under lib/db or lib/repositories in todox. Two of these rules exist because of a live SQL injection, not a hypothetical one.
datto-bcdr
Sync your whole Datto BCDR fleet into local SQLite and answer the questions the per-appliance Partner Portal can't: which backups failed screenshot verification, which are stale, and which clients are at risk. Trigger phrases: which datto backups failed screenshot verification, datto bcdr fleet health, which clients…
afi
The first CLI for Afi SaaS backup - full public-API coverage plus the fleet-wide coverage, staleness, and offboarding answers the rate-limited API can't serve live. Trigger phrases: check afi backup coverage, which mailboxes aren't backed up in afi, afi fleet health, offboard a user from afi backup, afi backup stale…
lastest
Run visual regression tests, review screenshot diffs, and manage baselines on a Lastest instance via the @lastest/mcp-server MCP tools.
lastest
Lastest is a visual regression testing platform. It records browser tests, runs them via Playwright, diffs screenshots, and uses AI to classify changes. Available as an MCP server for AI agent integration.