| Type | Skill | Skill | Skill |
|---|
| Categories | | | |
|---|
| In one line | An agentic skills framework and software methodology that works. | Garry Tan's Claude Code skill stack. | Production-grade engineering skills for 70+ coding agents. |
|---|
| Stage & funding | — | — | — |
|---|
| GitHub starsDiffers | ≈307k | ≈138k | ≈111k |
|---|
| License | MIT | MIT | MIT |
|---|
| LanguageDiffers | MDX | TypeScript | MDX |
|---|
| Last pushDiffers | Oct 23, 2026 | Oct 28, 2026 | Oct 13, 2026 |
|---|
| Works withDiffers | - Claude Code
- Codex
- Cursor
- opencode
- Gemini CLI
| | - Claude Code
- Cursor
- Copilot
- Any agent
|
|---|
| Built onDiffers | | | |
|---|
| Built on it | — | — | — |
|---|
| Similar toDiffers | | | |
|---|
| People linked | 1 | 1 | 1 |
|---|
| Events | 0 | 0 | 0 |
|---|
| Pros | - +DocsA full auto-triggering methodology: brainstorm, then a spec shown in digestible chunks, then a written plan, then subagent-driven development with review.[1]
- +DocsInstalls across many harnesses (Claude Code via the official plugin marketplace, Codex, Cursor, Gemini CLI, OpenCode, Copilot CLI and others).[1]
- +User reportsOne user reports being consistently happy with the brainstorming skill, and another says DeepSeek plus Superpowers in OpenCode feels on par with Opus.[1][2]
- +MetadataMIT, with about 308k stars and active releases (v7.0.1 on 2026-10-21).[1]
| - +DocsRole-based skills cover the sprint: /office-hours writes a design doc, /review hunts production bugs, /qa reproduces bugs in a real browser, /ship runs tests and opens the PR.[1]
- +DocsTeam mode bootstraps a repo so teammates get gstack automatically and stay on the same stable channel.[1]
- +DocsTelemetry is off by default and asked about once, with a documented privacy page.[1]
- +User reportsOne early user said it dramatically improved code quality and speed when used in Conductor.[1]
| - +Docs25 skills mapped to the lifecycle (spec, plan, build, test, review, ship) behind 9 slash commands, and they auto-activate by task.[1]
- +DocsInstalls into 70+ agents through the open skills CLI, either all at once or one skill at a time.[1]
- +User reportsOne user mentions it alongside Superpowers and Matt Pocock's skills as a methodology layer for larger projects.[1]
- +MetadataMIT and actively released (0.6.14 on 2026-10-29).[1]
|
|---|
| Cons | - −User reportsUsers report slower responses after installing and suspect skill context is filling the window (open, 27 comments).[1]
- −User reportsProcess friction: v7's brainstorm questions are called annoying, and plans over-specify implementation, leaving the executor no room for judgment.[1][2]
- −DocsBy default the brainstorming visual companion loads a logo from the vendor's site that reports your Superpowers version, which the README describes as telemetry.[1]
- −User reportsHeavyweight compared with minimal harnesses: users who like Pi cite Claude Code spinning up superpowers skills and many background agents.[1]
| - −DocsRequires Bun 1.4.2+ (plus Node.js on Windows), and every Claude Code session runs an auto-update check against GitHub (throttled to once an hour).[1]
- −User reportsNon-Claude hosts have been rough: '--host cursor' was rejected even though the README listed it, and the Conductor template failed on new projects (spring 2026, since closed).[1][2][3]
- −User reportsSome users worry its telemetry gives YC signal on what people build, even though telemetry is opt-in.[1][2]
| - −Docs/review collides with Claude Code's built-in command, so you have to use the namespaced /agent-skills:review.[1]
- −DocsThe optional '/build auto' implements every task after a single plan approval, removing human review between tasks (tests still run per task).[1]
- −MetadataLittle independent evaluation: its HN submission got 2 points and no comments, and skill packs like this are easy to copy.[1][2]
|
|---|
| Best for | Developers who want an opinionated spec-plan-execute discipline on top of any coding agent with little setup. | Founders and small teams on Claude Code who want opinionated reviewer, QA and release roles with minimal setup. | Developers who want senior-engineer workflows and quality gates packaged as portable skills across many agents. |
|---|
| Watch out for | Extra ceremony and context cost on small tasks; turn skills off where the process outweighs the change. | The suggested team-mode commit makes gstack 'required' for every teammate, so agree on that before committing it. | Watch for command-name clashes, and keep human review if you use /build auto. |
|---|
| Links | | | |
|---|