⏳ This skill is pending AI review.
Scores will appear once the review pipeline completes.
sw-autopilot
Use when the user wants several Superwiki tasks worked through one after another without being asked at each step (run the backlog, do all tasks of an area, keep going until done or blocked, act as orchestrator), or invokes sw-autopilot.
Choose how to use this skill
You do not need every option. Choose the path your AI client supports. The stable page stays the same; versioned files are immutable.
1. Native installer
This listing has no registered native installer command. Use the complete package or source fallback below, depending on what your client supports.
Do not guess an installer command or replace an existing version without reviewing the diff.
2. Complete package recommended
Download the ZIP when available. It includes SKILL.md plus the references, security notes and version metadata.
No complete ProSkills package is published for this listing yet.3. Prompt-only
Copy the prompt above when the agent can read the stable page or when you want to adopt the workflow without installing a skill.
Need only the instruction file?
Download SKILL.md only if your client requires a single file. The complete ZIP is safer for a full installation because it preserves the references and release context.
No path installs or executes anything by itself. Your agent still needs access to the project files. Before updating, compare the installed version and review the diff.
// RATINGS
// README
Superwiki
Agent skills that turn a project's docs/ folder into an LLM-maintained wiki and task tracker. Your coding agent writes it and keeps it current; you read it as an Obsidian vault or in a built-in viewer. Works with Claude Code, Codex CLI and GitHub Copilot CLI.
Live demo · Install · Use · Design notes
Quick start
Requires Node 18 or newer.
npx superwiki install claude # or: codex, copilot, all
Start a new agent session in your project and run /sw-init. It sets up docs/ and asks whether you want the task tracker. From then on, ask for what you need in plain words, or use a command.
What you get
docs/
index.md the open tasks, then the catalog of the wiki, one line per page
log.md append-only history
raw/ your sources, never modified
wiki/ pages the agent writes
tasks/ one file per task (optional)
plans/ one plan per task
viewer.html the task board and the wiki in a browser
sw-init also adds a short block of rules to AGENTS.md, so the agent maintains the vault in every session, with or without a command.
Everything is plain markdown. docs/ keeps working as an Obsidian vault without Superwiki.
The wiki
Superwiki follows the LLM Wiki pattern described by Andrej Karpathy: you curate raw sources, the agent owns the wiki, and a short set of rules tells it how to maintain it. Hand the agent an article, a transcript or your notes, and it files a summary and links it into the catalog.
The tasks
A task is one file with a goal, a "Done when" list and its dependencies. The agent will not start a task whose dependencies are open. A task is done only when every "Done when" item has been checked by a command run in the closing session, one that would have failed if the item did not hold. The evidence is written into the task file, so you can read later how each item was proven.
index.md opens with the task list, so you can follow the work without the viewer:
## Tasks
ready 9 · in progress 1 · blocked 18 · done 20
**In progress**
- [[B-20]] Portfolio sync · M5
**Ready**
- [[F-01]] Frontend skeleton and guards · M0
- [[B-18]] Valuation · M4
**Blocked**
- [[B-21]] Journal and thesis gates · M5 · waits on B-20
**Done (20)** [[B-01]] [[B-02]] [[B-03]] ...
The list is generated from the task files. Do not edit it by hand: change the task file, and the agent (or /sw-index) rewrites the list.
Built to be cheap for the agent
- Little to read. One small index, one file per task, and a script that answers "what is ready?" or "what blocks this?" without the agent reading the vault.
- Work in clean contexts. Planning, implementing and reviewing run in subagents, each on the model you choose. The main session only tracks the task's status and checks the result, so it stays small.
- Cost you can see.
sw-usageshows what a session used, per agent;sw-doctorshows what every session carries before it starts.
Measured on a real project with 165 tasks, converted from a single markdown index:
| Before | After | |
|---|---|---|
| Read at the start of every session | 197 KB index | 7.7 KB index + 2.7 KB of rules |
| Read to start one task | the index, then the task's section | one file, 2 KB at the median |
| Marking a task done | a status cell, plus a ✅ at every reference to it (median 12 places) | one frontmatter line |
A pilot run on 2026-10-06 compared Superwiki with a single-session workflow on two backend tasks of one project, two runs each, with a blind review of the results. On the small task Superwiki cost $2.79 and $3.23 against $4.18 and $4.81. On the medium task the two overlapped, and the review scored quality a tie. Two tasks, one project, one day: read it as a direction. DESIGN.md has the figures and the limits.
Status: early. Not every skill has been run in every agent. DESIGN.md lists what has been proven and what has not.
Install
npx superwiki install <target>...
| Target | Agent | What is installed | Invoke a skill with |
|---|---|---|---|
claude | Claude Code | a copy of each skill | /sw-<name>, or the skill named in a sentence |
codex | Codex CLI | a copy of each skill | $sw-<name>, or the skill named in a sentence |
copilot | GitHub Copilot CLI | a copy of each skill | /sw-<name>, or the skill named in a sentence |
all | the three above |
The skills are named sw-<name> in every agent: sw-plan, sw-implement, sw-plan-implement. Each is a folder of that name, and every install is a copy of those folders, or a link to them.
If you installed an earlier version for Claude Code, install claude also removes its plugin sw@superwiki and the marketplace superwiki at the same scope, so the skills are not listed twice. That needs the claude command on your PATH; without it, remove them with claude plugin uninstall sw@superwiki and claude plugin marketplace remove superwiki.
Start a new agent session after installing: a running session does not pick up new skills.
Home folder or project
The installer asks where the skills should go:
- Your home folder. They serve every project on this machine. This is the usual choice.
- This project, the directory you run the command in. They are committed with the repository, so teammates and cloud agents have them too.
A flag answers in advance. Without a terminal (a script, CI) nothing is asked and the home folder is used.
npx superwiki install --global claude # home folder, no question
npx superwiki install --project . claude codex # this project, no question
| Target | In your home folder | In a project |
|---|---|---|
claude | ~/.claude/skills | .claude/skills |
codex | ~/.agents/skills | .agents/skills |
copilot | ~/.copilot/skills | .agents/skills |
A skill folder of the same name that Superwiki did not install is kept and reported; --force replaces it.
Skills kept in a project are markdown and scripts inside your repository, so the project's own linter and formatter may pick them up. If they report on .claude/skills or .agents/skills, add those folders to their ignore lists.
Cloud agents
A cloud agent (Claude Code on the web, Codex cloud, the GitHub Copilot coding agent) starts from a clone of your repository and never sees your home folder. It has the Superwiki skills only if they are in the repository.
- Put the skills in the project:
npx superwiki install --project . claude codex, or answer "this project" in the installer. On a new vault,/sw-initalso asks whether to keep the skills in the repository, and for which agents. - Commit and push
.claude/skillsand.agents/skills.
| Cloud agent | Target | Reads the skills from |
|---|---|---|
| Claude Code on the web | claude | .claude/skills |
| Codex cloud | codex | .agents/skills |
| GitHub Copilot coding agent | copilot | .agents/skills |
These places are the ones each tool's documentation names. Superwiki has not yet been run in a cloud session.
Other agents
Agents that load SKILL.md folders from ~/.agents/skills get the skills with the target global. For an agent with a skills folder of its own (Cursor, Gemini CLI, OpenCode and others), copy each skills/sw-<name> folder from a clone into it, unchanged. Neither has been tested. Expect these limits:
- planning, implementing and reviewing run in the main session, not in subagents;
- `s
// HOW IT'S BUILT
KEY FILES