Ports the obra/superpowers methodology as 14 native DSH skills covering planning, TDD, debugging, and code review.
Install
# from npm (prebuilt)
dsh plugin --profile web add @jasonfreelab/dsh-superpowers
# from GitHub (first run asks for allowBuilds approval — follow the hint, retry)
dsh plugin --profile web add github:JasonFreeLab/dsh-superpowers
Any plugin you install runs third-party code with your own permissions — it can read your files, use your credentials, and reach the network, and tool approvals don’t sandbox it. GitHub-sourced plugins also run build scripts at install time — pnpm blocks those until you allow them, so an install can stop with ERR_PNPM_GIT_DEP_PREPARE_NOT_ALLOWED or ERR_PNPM_IGNORED_BUILDS; dsh prints the exact key to add under allowBuilds in your profile’s pnpm-workspace.yaml, and the install works on the next run. Allowing a build is a trust decision: only install sources you trust, and pin a commit (github:owner/repo#sha).
README
English | 中文
A DSH (DeepSeek Harness) port of obra/superpowers — the full multi-agent software-development methodology, available out of the box as native DSH skills.
Ported from upstream obra/superpowers (v6.4.2, by Jesse Vincent / Prime Radiant). Skill content is taken directly from upstream and mapped onto the DSH toolset.
Table of Contents
Features
- 15 methodology skills — brainstorm → plan → TDD → systematic debugging → code review → integrate, plus session diagnosis.
- Native DSH skills — injected into
ctx.skillsvia aSkillProvider(rank 550; overridable by project/user skills). - English (i18n) — skill content kept in original English; Chinese docs in
README.zh.md. - Zero-build install —
lib/is committed, so GitHub installs need no build step. - Automated releases —
release-please(versioning + CHANGELOG + release notes) plus trusted publishing to npm and GitHub Packages.
Install
Using the web profile as the example (change --profile for others). Requires Node >= 20, pnpm >= 9, and dsh.
# From npm (recommended)
dsh plugin --profile web add @jasonfreelab/dsh-superpowers
# Local path install (dev / offline)
git clone https://github.com/JasonFreeLab/dsh-superpowers.git && cd dsh-superpowers
pnpm install && pnpm build
dsh plugin --profile web add ./
# Verify: you should see id: superpowers
dsh --profile web --dump-config | grep -A2 '@jasonfreelab/dsh-superpowers'
# Update / uninstall
dsh plugin --profile web add @jasonfreelab/dsh-superpowers
dsh plugin --profile web remove @jasonfreelab/dsh-superpowers
Also published to GitHub Packages (requires auth):
npm.pkg.github.com/@jasonfreelab/dsh-superpowers.
Usage
In the DSH web UI
- Start the web UI and open the printed URL:
dsh web(alias ofdsh --profile web). - Start a new session. The 15 skills are registered automatically in
ctx.skillsand appear in the model's<available_skills>catalog — no extra setup. - The model loads the matching skill by itself via the
skilltool. - You can also invoke a skill explicitly with
/skill-name, e.g./superpower-brainstorming.
Typical flows
Build me X → superpower-brainstorming → superpower-writing-plans → superpower-subagent-driven-development
Fix this bug → superpower-systematic-debugging
Review this → superpower-requesting-code-review
Session went wrong → superpower-diagnosing-superpowers
Verify: inside a session await ctx.skills.list({cwd}) should return 15 entries with provider: superpowers.
Included skills
| Skill | When it triggers |
|---|---|
superpower-using-superpowers |
Use when starting any conversation — establishes how to find and use skills, requiring skill invocation before ANY response including clarifying questions |
superpower-brainstorming |
Use before any creative work — creating features, building components, adding functionality, or modifying behavior |
superpower-writing-plans |
Use when you have a spec or requirements for a multi-step task, before touching code |
superpower-using-git-worktrees |
Use when starting feature work that needs isolation from the current workspace, or before executing implementation plans |
superpower-executing-plans |
Use when executing a plan inline in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available |
superpower-subagent-driven-development |
Use when executing implementation plans with independent tasks in the current session |
superpower-dispatching-parallel-agents |
Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies |
superpower-test-driven-development |
Use when implementing any feature or bugfix, before writing implementation code |
superpower-systematic-debugging |
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes |
superpower-verification-before-completion |
Use before claiming work is complete, fixed, or passing — run verification commands and confirm output before any success claim |
superpower-requesting-code-review |
Use when completing tasks, implementing major features, or before merging |
superpower-receiving-code-review |
Use when receiving code review feedback, before implementing suggestions |
superpower-finishing-a-development-branch |
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work |
superpower-writing-skills |
Use when creating new skills, editing existing skills, or verifying skills work before deployment |
superpower-diagnosing-superpowers |
Use when a superpowers session went wrong (repeated work, ignored plans, stumbles, a skill that didn't fire, unexpected cost) — investigate it, or build a maintainer bug report |
Tool mapping
Upstream references Claude Code tools; this package maps them onto DSH tools (see skills/superpower-using-superpowers/references/dsh-tools.md):
| Claude Code | DSH |
|---|---|
| Bash | bash |
| Read / Write / Edit | read / write / edit |
| Glob / Grep | glob / grep |
| TodoWrite | todo_write |
| Task (subagent) | subagent / subagent_fork |
| ExitPlanMode | exit_plan_mode |
| AskUserQuestion | ask_user_question |
| WebFetch / WebSearch | web_fetch / web_search |
| Present a deliverable | present |
| Read an image | read_image |
| Load a skill | skill tool / /name gesture |
Layout
src/superpowers.ts # SkillProvider (rank 550), lazily loads SKILL.md bodies
skills/ # 15 skills (English, with references/)
lib/ # build output (committed for zero-build installs)
scripts/ # test.mjs + verify.mjs
cordis.patch.yml # bundle patch
.github/workflows/ # ci.yml + release.yml + release-please.yml
Development
pnpm install
npm run build # or: pnpm build
npm run typecheck
npm test # comprehensive (content hygiene + real SkillRegistry)
npm run verify # structural + runtime smoke
Contributing
Issues and pull requests are welcome. The upstream methodology lives in obra/superpowers.
License
MIT, matching upstream obra/superpowers (Jesse Vincent / Prime Radiant).
Links
More in this category
zhu1090093659/dsh-web#packages/dsh-skill-explorer★ 8707
Skill center for the dsh web GUI: browse all loaded skills grouped by source, enable or disable model invocation, create new skills, and delete into a recoverable trash.
GanyuanRan/Aegis★ 1346
Software-engineering method pack for coding agents, with skills for baseline-first planning, systematic debugging, prompt hygiene, verification before completion, and repair/retirement tracking.
superdesigndev/superdesign-skill★ 666
Design skill for UI and marketing graphics on the Superdesign canvas: reads the repo for context, extracts its design system, then generates and iterates branchable design drafts, flow pages, and reusable components through the Superdesign CLI.
linhay/harmony-next.skills★ 362
HarmonyOS NEXT skill bundle for DeepSeek Harness with offline API references and DevEco, HDC, and emulator automation guidance.
dhicoc/dsh-reverse-skill★ 246
Complete reverse-skill pack (85 SKILL.md) as a DeepSeek Harness Cordis plugin: reverse engineering, authorized pentesting and security-research skill router.
sandbaseai/sandbase-skills★ 202
Mounts 88 packaged research, social-intelligence, marketing and business Agent Skills into dsh through the filesystem Skill provider.
Community comments
Comments are public GitHub Discussions. Loading them connects to GitHub and Giscus; a GitHub account is required to post.