# superpowers: what changed from 5 to 6

- Product: superpowers (https://whatsnew.fyi/product/superpowers)
- Vendor: obra
- Range: changelog entries numbered after v5.1.0 up to and including v6.3.0, stable releases only
- Entries below: 7 releases (newest first)
- Resolved: 5 is v5.1.0 and 6 is v6.3.0, the newest stable release of each major we track
- Carrying security changes: 0 · CVEs mentioned: 0 · Mentioning breaking changes: 0 · Removing or deprecating something: 3
- Page: https://whatsnew.fyi/product/superpowers/compare/5...6

What's New is an index, not a publisher: every entry below links to the vendor's own release notes, which are the authoritative source. Entries are labelled where they are hand-curated sample data, pre-releases, or drawn from a secondary source such as a developer blog.

Reuse: the summaries, labels and curation here are © What's New. Quote freely with attribution and a link back; wholesale republication of the corpus is not permitted — terms: https://whatsnew.fyi/terms. The vendors' own release notes remain their publishers'.

## What changed (61 changes, grouped by kind)

### Added

#### v6.3.0 (2026-08-12)

- Devin CLI support to install superpowers via `devin plugins install obra/superpowers` with skills auto-triggering at session start
- Hermes Agent support to install from git clone with skills registering to Hermes' native loader and bootstrap loading on first turn
- Grok Build CLI added to install documentation

#### v6.2.0 (2026-07-24)

- Gemini CLI support is restored with install docs and the gemini-tools.md tool-mapping reference

#### v6.1.1 (2026-07-02)

- Add package-codex-plugin.sh script for building deterministic Codex portal packages with timestamp normalization, executable mode preservation, OpenAI metadata verification, and dirty worktree detection

#### v6.1.0 (2026-06-30)

- Codex can now install from the marketplace with a repo-local Codex marketplace manifest at .agents/plugins/marketplace.json

#### v6.0.3 (2026-06-18)

- sdd-workspace helper to resolve SDD workspace per worktree

#### v6.0.0 (2026-06-16)

- Add support for Kimi Code harness with plugin manifest and marketplace installation
- Add support for Pi harness as a session-start extension with native skills
- Add support for Antigravity harness with direct plugin installation and bootstrap from first message
- Implement pre-flight plan validation to check for internal conflicts before the first task
- Add task-brief and review-package scripts to write task text and review diffs as files instead of pasted text
- Add Global Constraints block to plans to list rules that bind every task
- Add per-task Interfaces block to plans to specify what each task consumes and produces
- Implement single whole-branch review at end of run on the most capable model instead of re-reviewing task by task

### Changed

#### v6.3.0 (2026-08-12)

- Ceremony now scales to the task by classifying requests as spike, bounded, or architectural, with small tasks skipping the two-document ritual
- Controllers no longer stall on plan conflicts; non-catastrophic conflicts get a recorded ruling and work continues while only destructive or irreversible actions stop for human approval
- Pre-dispatch conflict scan records its checks in the ledger instead of just asserting the plan is clean
- Small same-shape tasks batch into one dispatch, cutting subagent cost on micro-task plans
- Implementers and reviewers may not spawn their own subagents to prevent duplicate reviews
- Plans carry a Spec pointer and SDD reads the spec at setup so plan conflicts get resolved against the design
- Reviewers re-read evidence they find illegible instead of re-running the test suite
- Subagent waits are event-driven instead of poll-heavy, spawn pin model and reasoning effort explicitly
- Worktree removal no longer destroys untracked files; when git worktree remove refuses due to uncommitted work, the skill stops and asks instead of using force
- README reorganized with added table of contents and restructured Getting Started section

#### v6.2.0 (2026-07-24)

- The workspace is now plan-scoped: .superpowers/sdd/ resolves to a per-plan directory .superpowers/sdd/<plan-basename>/, sdd-workspace requires the plan file, task-brief and review-package write into their plan's directory, the ledger names its plan on its first line, and the workspace is deleted once the final review is clean
- The review-fix loop now resumes the implementer with resume-the-implementer semantics instead of fresh dispatches, includes a scoped re-review prompt so the re-reviewer checks the fixes rather than re-reading the whole task, and installs a five-round circuit breaker with controller adjudication when it trips
- testing-anti-patterns.md is renamed to writing-good-tests.md and rebuilt as a positive catalog with six rules, a falsifiability discipline including naming production changes that would fail the test, and a closing mutation check
- TDD's Why Order Matters section is removed but its prose rebuttals survive as rationalization rows in the Common Rationalizations table
- finishing-a-development-branch no longer offers to discard work in the completion menu; discard survives only as an explicit-request-only path with typed-confirmation ritual
- Recap and persuasion prose removed from brainstorming, systematic-debugging, dispatching-parallel-agents, verification-before-completion, executing-plans, subagent-driven-development, requesting-code-review, receiving-code-review, using-git-worktrees, writing-plans, and writing-skills
- PR creation is now forge-agnostic, using your forge's CLI or the URL printed on push instead of a blessed list of tools
- The SessionStart hook now dispatches via Git Bash by declaring shell: bash, which Claude Code ≥ 2.1.81 resolves to Git for Windows directly and surfaces an actionable install prompt when Git Bash is missing

#### v6.1.1 (2026-07-02)

- Move worked shell-hook example in docs/porting-to-a-new-harness.md from Codex to Cursor

#### v6.1.0 (2026-06-30)

- Compressed the using-superpowers bootstrap by replacing the graphviz skill-flow diagram with prose, folding the Instruction-Priority section into User Instructions, dropping the per-platform How to Access Skills walkthrough, and trimming the Platform Adaptation pointer
- Pruned per-harness tool-mapping references to retain only harness-specific notes such as subagent dispatch, task tracking, and instructions-file paths

#### v6.0.3 (2026-06-18)

- SDD scratch files moved from .git/ to .superpowers/sdd/ directory to avoid Claude Code protection conflicts
- .superpowers/sdd/ directory is git-ignored and kept out of git status and commits

#### v6.0.0 (2026-06-16)

- Rewrite of subagent-driven-development task review system to use a single reviewer per task instead of two, reducing token usage by approximately 50% and improving speed
- Replace separate spec-reviewer-prompt.md and code-quality-reviewer-prompt.md files with a single task-reviewer-prompt.md file
- Move worktrees from global ~/.config/superpowers/worktrees/ directory to project-local .worktrees/ directory
- Make reviewer prompts read-only and prevent reviewers from modifying the working tree or branch
- Require explicit model specification for every dispatch instead of allowing unnamed models to inherit the session's most expensive model
- Ban controller ability to suppress findings or pre-rate severity in reviewer judgments
- Add file and line evidence backing to reviewer answers and progress ledger for resuming interrupted work
- Rewrite skills' tool calls to be significantly more vendor-neutral
- Improve brainstorming visual companion security model with authentication
- Improve brainstorming visual companion to survive restarts and dropped connections

### Fixed

#### v6.3.0 (2026-08-12)

- render-graphs.js in writing-skills now works on Windows
- Corrected Copilot CLI backgrounding guidance for Windows
- bump-version.sh now covers the Hermes manifest

#### v6.2.0 (2026-07-24)

- find-polluster.sh now correctly matches test files by fixing the ./ prefix mismatch in the -path pattern, preventing double-prefixing when a caller-supplied pattern includes the prefix, and handling **/ patterns that were silently skipping tests directly under the base directory
- The worktree path in finishing-a-development-branch is no longer recomputed after cleanup has changed directory, fixing a bug where provenance checks never matched and cleanup silently no-oped

#### v6.1.1 (2026-07-02)

- Codex no longer re-registers the Claude SessionStart hook by declaring an explicit empty hooks object in the manifest
- Correct stale hooks-codex.json pointer in docs/windows/polyglot-hooks.md
- Fix the Codex plugin category to Developer Tools

### Removed

#### v6.1.1 (2026-07-02)

- Remove orphaned Codex session-start dead code in hooks/session-start-codex and its test cases

#### v6.1.0 (2026-06-30)

- Deleted claude-code-tools.md and copilot-tools.md reference files as they contained no harness-specific content
- Codex no longer ships a SessionStart hook; removed hooks-codex.json config and its manifest registration
- Removed Gemini CLI support following Google's EOL on 2026-06-18; Gemini is removed from install docs, subagent-capable platform lists, eval-harness description, and its tool-mapping reference is deleted

#### v6.0.2 (2026-06-17)

- Remove the evals submodule from the shipped plugin as it broke plugin installs for some users

## Release notes

### v6.3.0

- Date: 2026-08-12
- Version: v6.3.0
- Original notes: https://github.com/obra/superpowers/releases/tag/v6.3.0
- Permalink: https://whatsnew.fyi/product/superpowers/releases/v6.3.0

- **added** — Devin CLI support to install superpowers via `devin plugins install obra/superpowers` with skills auto-triggering at session start
- **added** — Hermes Agent support to install from git clone with skills registering to Hermes' native loader and bootstrap loading on first turn
- **added** — Grok Build CLI added to install documentation
- **changed** — Ceremony now scales to the task by classifying requests as spike, bounded, or architectural, with small tasks skipping the two-document ritual
- **changed** — Controllers no longer stall on plan conflicts; non-catastrophic conflicts get a recorded ruling and work continues while only destructive or irreversible actions stop for human approval
- **changed** — Pre-dispatch conflict scan records its checks in the ledger instead of just asserting the plan is clean
- **changed** — Small same-shape tasks batch into one dispatch, cutting subagent cost on micro-task plans
- **changed** — Implementers and reviewers may not spawn their own subagents to prevent duplicate reviews
- **changed** — Plans carry a Spec pointer and SDD reads the spec at setup so plan conflicts get resolved against the design
- **changed** — Reviewers re-read evidence they find illegible instead of re-running the test suite
- **changed** — Subagent waits are event-driven instead of poll-heavy, spawn pin model and reasoning effort explicitly
- **changed** — Worktree removal no longer destroys untracked files; when git worktree remove refuses due to uncommitted work, the skill stops and asks instead of using force
- **fixed** — render-graphs.js in writing-skills now works on Windows
- **fixed** — Corrected Copilot CLI backgrounding guidance for Windows
- **fixed** — bump-version.sh now covers the Hermes manifest
- **changed** — README reorganized with added table of contents and restructured Getting Started section

###### Harness Support

- **Devin CLI**: `devin plugins install obra/superpowers` now works, and skills auto-trigger at session start. (#1995)
- **Hermes Agent**: install from a git clone; skills register with Hermes' native loader and the bootstrap loads on the first turn. (#1922, #2025)
- **Grok Build CLI** added to the install docs. (#1919)

###### Brainstorming

- **Ceremony now scales to the task.** Requests are classified as spike, bounded, or architectural; small tasks skip the two-document ritual. Every path still stops for your approval before implementation. (#2063)

###### Subagent-Driven Development

- **Controllers no longer stall on plan conflicts.** Non-catastrophic conflicts and ambiguities get a recorded ruling and work continues; only destructive or irreversible actions still stop for a human. One donated session had sat blocked for almost nine hours on a question the controller could have decided. (#2077)
- **The pre-dispatch conflict scan records its checks in the ledger** instead of just asserting the plan is clean. (#2080)
- **Small same-shape tasks batch into one dispatch**, cutting subagent cost sharply on micro-task plans; batch reviews verify every file in the brief made it into the diff. (#2078)
- **Implementers and reviewers may not spawn their own subagents**, which was producing duplicate reviews. (#2059)
- **Plans carry a `Spec:` pointer** and SDD reads the spec at setup, so plan conflicts get resolved against the design instead of guessed at. (#2086)
- Reviewers re-read evidence they find illegible instead of re-running the test suite (#2089), and circuit-breaker rulings now show up in the Finish report.

###### Codex

- Subagent waits are event-driven instead of poll-heavy, spawns pin model and reasoning effort explicitly, and the multi-agent reference is corrected against Codex source. (#2060, #2061, #2062)

###### Finishing a Development Branch

- **Worktree removal no longer destroys untracked files.** When `git worktree remove` refuses because the tree holds uncommitted work, the skill stops, names the files, and asks — instead of reaching for `--force`. (#2016, #1223, #2024)

###### Fixes

- `render-graphs.js` in writing-skills works on Windows.
- Corrected Copilot CLI backgrounding guidance for Windows. (#1929, #2006)
- `bump-version.sh` covers the Hermes manifest.

###### Documentation

- README: added a table of contents and reorganized Getting Started.

### v6.2.0

- Date: 2026-07-24
- Version: v6.2.0
- Original notes: https://github.com/obra/superpowers/releases/tag/v6.2.0
- Permalink: https://whatsnew.fyi/product/superpowers/releases/v6.2.0

- **changed** — The workspace is now plan-scoped: .superpowers/sdd/ resolves to a per-plan directory .superpowers/sdd/<plan-basename>/, sdd-workspace requires the plan file, task-brief and review-package write into their plan's directory, the ledger names its plan on its first line, and the workspace is deleted once the final review is clean
- **changed** — The review-fix loop now resumes the implementer with resume-the-implementer semantics instead of fresh dispatches, includes a scoped re-review prompt so the re-reviewer checks the fixes rather than re-reading the whole task, and installs a five-round circuit breaker with controller adjudication when it trips
- **changed** — testing-anti-patterns.md is renamed to writing-good-tests.md and rebuilt as a positive catalog with six rules, a falsifiability discipline including naming production changes that would fail the test, and a closing mutation check
- **changed** — TDD's Why Order Matters section is removed but its prose rebuttals survive as rationalization rows in the Common Rationalizations table
- **changed** — finishing-a-development-branch no longer offers to discard work in the completion menu; discard survives only as an explicit-request-only path with typed-confirmation ritual
- **changed** — Recap and persuasion prose removed from brainstorming, systematic-debugging, dispatching-parallel-agents, verification-before-completion, executing-plans, subagent-driven-development, requesting-code-review, receiving-code-review, using-git-worktrees, writing-plans, and writing-skills
- **changed** — PR creation is now forge-agnostic, using your forge's CLI or the URL printed on push instead of a blessed list of tools
- **changed** — The SessionStart hook now dispatches via Git Bash by declaring shell: bash, which Claude Code ≥ 2.1.81 resolves to Git for Windows directly and surfaces an actionable install prompt when Git Bash is missing
- **added** — Gemini CLI support is restored with install docs and the gemini-tools.md tool-mapping reference
- **fixed** — find-polluster.sh now correctly matches test files by fixing the ./ prefix mismatch in the -path pattern, preventing double-prefixing when a caller-supplied pattern includes the prefix, and handling **/ patterns that were silently skipping tests directly under the base directory
- **fixed** — The worktree path in finishing-a-development-branch is no longer recomputed after cleanup has changed directory, fixing a bug where provenance checks never matched and cleanup silently no-oped

##### v6.2.0 (2026-07-23)

###### Subagent-Driven Development

Two structural changes to how SDD tracks progress and closes out review findings, both developed against live eval campaigns.

- **The workspace is now plan-scoped.** `.superpowers/sdd/` had no plan identity and no end-of-life: a follow-up plan in the same working tree could read the previous plan's ledger as its own progress (observed in the wild, with multiple contamination rounds and ad-hoc workarounds). `sdd-workspace` now requires the plan file and resolves a per-plan directory, `.superpowers/sdd/<plan-basename>/`; `task-brief` and `review-package` write into their plan's directory (`review-package` gains the plan file as its first argument); the ledger names its plan on its first line; and the workspace is deleted once the final review is clean — git history is the durable record. Baseline evals showed controllers already refused foreign ledgers, but at a cost of 6–13 tool calls of cross-plan git forensics per resume; plan-scoping makes the answer structural instead. (25/25 baseline and GREEN eval runs documented in `docs/specs/` and `docs/plans/`.)
- **The review-fix loop resumes the implementer.** The lifecycle restructure gives fix rounds resume-the-implementer semantics instead of fresh dispatches, adds a scoped re-review prompt (`re-review-prompt.md`) so the re-reviewer checks the fixes rather than re-reading the whole task, and installs a five-round circuit breaker with controller adjudication when it trips. SKILL.md reorganizes by lifecycle, and its Red Flags convert to the house rationalization-table form.

###### Skills

A branch-wide compression campaign: recap sections, social proof, and benefits-selling prose aimed at a reader who has already invoked the skill are gone, with every load-bearing argument folded into a rationalization-table row or moved to its point of use. Each cut was micro-tested with subagent probes, and the one cut that measurably degraded behavior was reworked rather than shipped.

- **`testing-anti-patterns.md` is now `writing-good-tests.md`.** The TDD reference doc is rebuilt as a positive catalog — six rules that lead with the GOOD example — and absorbs a falsifiability discipline: name the production change that would fail the test, derive expectations independently of the code under test, and a closing mutation check. It closes two holes by name: the string-presence trap (grep-style tests on scripts, skills, and prompts counterfeit falsifiability — the observable is behavior, never text) and the change-detector trap (a constant assertion can fail and still protect nothing), each with a hard stop in the gate function. Trivial code and human prose earn no test; the trigger broadens from "adding mocks" to any test writing.
- **TDD's "Why Order Matters" rebuttals survive as rationalization rows.** Deleting the section outright measurably degraded test-first behavior under "just write it, tests after" pressure (control 8/10 → treatment 5/10, corroborated on Claude and Codex), so each prose rebuttal now lives in its Common Rationalizations row — the section is gone but the arguments fire where an agent hits them mid-rationalization.
- **`finishing-a-development-branch` no longer offers to discard your work.** The completion menu dates from when throwing away branches was routine; "Discard this work" next to "Merge" advertised destroying finished, passing work. Discard survives as an explicit-request-only path with the same typed-confirmation ritual. The same pass made PR creation forge-agnostic (your forge's CLI or the URL printed on push, not a blessed list of tools) and fixed a real bug: the worktree path was recomputed after cleanup had already changed directory, so provenance checks never matched and cleanup silently no-oped.
- **Recap and persuasion prose removed across the library.** `brainstorming`, `systematic-debugging`, `dispatching-parallel-agents`, `verification-before-completion`, `executing-plans`, `subagent-driven

_[Truncated at 4000 characters — full notes: https://github.com/obra/superpowers/releases/tag/v6.2.0]_

### v6.1.1

- Date: 2026-07-02
- Version: v6.1.1
- Original notes: https://github.com/obra/superpowers/releases/tag/v6.1.1
- Permalink: https://whatsnew.fyi/product/superpowers/releases/v6.1.1

- **fixed** — Codex no longer re-registers the Claude SessionStart hook by declaring an explicit empty hooks object in the manifest
- **removed** — Remove orphaned Codex session-start dead code in hooks/session-start-codex and its test cases
- **changed** — Move worked shell-hook example in docs/porting-to-a-new-harness.md from Codex to Cursor
- **fixed** — Correct stale hooks-codex.json pointer in docs/windows/polyglot-hooks.md
- **fixed** — Fix the Codex plugin category to Developer Tools
- **added** — Add package-codex-plugin.sh script for building deterministic Codex portal packages with timestamp normalization, executable mode preservation, OpenAI metadata verification, and dirty worktree detection

##### v6.1.1 (2026-07-02)

###### Codex

- **Codex no longer re-registers the Claude SessionStart hook.** v6.1.0 removed the Codex hook config and its manifest `hooks` pointer, meaning to stop Codex from installing a SessionStart hook — but with no `hooks` field, Codex fell back to auto-discovering `hooks/hooks.json`, the Claude Code SessionStart hook that the marketplace ships from the repo root, and re-registered it along with its install-time trust prompt. The Codex manifest now declares an explicit empty hooks object (`hooks: {}`), which Codex reads as "no hooks" instead of reaching the auto-discovery fallback. An absent field, `[]`, and an empty inline list all collapse back to the fallback, so the value has to be exactly `{}`.
- **Removed orphaned Codex session-start dead code.** `hooks/session-start-codex` had no caller once the Codex hook config was deleted, so it and its redundant test cases are gone. The worked shell-hook example in `docs/porting-to-a-new-harness.md` moves from Codex — now native skill discovery with no session-start hook — to Cursor, a live shell-hook harness, and the stale `hooks-codex.json` pointer in `docs/windows/polyglot-hooks.md` is corrected. The Codex plugin category is also fixed to "Developer Tools".

###### Packaging

- **New `package-codex-plugin.sh` for building the Codex portal package.** A maintainer script produces a deterministic Codex "portal" archive — `.zip` by default, `tar.gz` on request — that normalizes entry timestamps, preserves executable modes, verifies every packaged skill ships its OpenAI metadata, includes the app and composer icons, and refuses to run against a dirty worktree. The packaged manifest keeps the source `hooks: {}` object so a portal-installed plugin avoids the same SessionStart auto-discovery, and the script can rebuild a byte-identical archive from a saved metadata source. Covered by a new test suite.

### v6.1.0

- Date: 2026-06-30
- Version: v6.1.0
- Original notes: https://github.com/obra/superpowers/releases/tag/v6.1.0
- Permalink: https://whatsnew.fyi/product/superpowers/releases/v6.1.0

- **changed** — Compressed the using-superpowers bootstrap by replacing the graphviz skill-flow diagram with prose, folding the Instruction-Priority section into User Instructions, dropping the per-platform How to Access Skills walkthrough, and trimming the Platform Adaptation pointer
- **changed** — Pruned per-harness tool-mapping references to retain only harness-specific notes such as subagent dispatch, task tracking, and instructions-file paths
- **removed** — Deleted claude-code-tools.md and copilot-tools.md reference files as they contained no harness-specific content
- **added** — Codex can now install from the marketplace with a repo-local Codex marketplace manifest at .agents/plugins/marketplace.json
- **removed** — Codex no longer ships a SessionStart hook; removed hooks-codex.json config and its manifest registration
- **removed** — Removed Gemini CLI support following Google's EOL on 2026-06-18; Gemini is removed from install docs, subagent-capable platform lists, eval-harness description, and its tool-mapping reference is deleted

###### Lower Per-Session Token Cost

The `using-superpowers` bootstrap is injected into every session, so its size is paid for constantly. This release trims it and the per-harness references it points to, without dropping behavior-shaping content.

- **Compressed the `using-superpowers` bootstrap.** Replaced the graphviz skill-flow diagram with the prose it encoded, folded the standalone Instruction-Priority section into User Instructions, dropped the per-platform "How to Access Skills" walkthrough, and trimmed the Platform Adaptation pointer to the harnesses that still ship a reference file. The full Red Flags rationalization table and the user-instruction precedence rules are unchanged.
- **Pruned the per-harness tool-mapping references.** The verbose action-to-tool tables restated guidance modern agents already follow. Each reference file is trimmed to the harness-specific notes that still carry weight — subagent dispatch, task tracking, instructions-file paths — and `claude-code-tools.md` and `copilot-tools.md`, which had nothing harness-specific left, are deleted.

###### Codex

- **Codex can install from the marketplace.** Codex marketplace sources expect a `.agents/plugins/marketplace.json` at the marketplace root; the repo only shipped the Claude marketplace file, so Codex could name the marketplace but found no installable plugin entries. A repo-local Codex marketplace manifest now points at the same repository root, so the plugin is installable from Codex.
- **Codex no longer ships a SessionStart hook.** Codex reliably triggers skills on its own, and the bootstrap hook made the UX worse rather than better. The Codex hook config (`hooks-codex.json`) and its manifest registration are removed.

###### Harness Support

- **Gemini CLI support removed.** Google EOLed the Gemini CLI on 2026-06-18; the extension can no longer be installed or updated. Gemini is gone from the install docs, the subagent-capable platform lists, and the eval-harness description, and its tool-mapping reference is deleted.

### v6.0.3

- Date: 2026-06-18
- Version: v6.0.3
- Original notes: https://github.com/obra/superpowers/releases/tag/v6.0.3
- Permalink: https://whatsnew.fyi/product/superpowers/releases/v6.0.3

- **changed** — SDD scratch files moved from .git/ to .superpowers/sdd/ directory to avoid Claude Code protection conflicts
- **added** — sdd-workspace helper to resolve SDD workspace per worktree
- **changed** — .superpowers/sdd/ directory is git-ignored and kept out of git status and commits

###### Subagent-Driven Development

- **SDD scratch files moved out of `.git/`.** Claude Code treats `.git/` as a protected path and denies agent writes there, so an implementer subagent writing its report into `.git/sdd/` got blocked mid-run. Task briefs, implementer reports, review diffs, and the progress ledger now live in a self-ignoring `.superpowers/sdd/` directory in the working tree — kept out of `git status` and out of commits, and resolved per worktree by a shared `sdd-workspace` helper. One caveat: because the workspace is git-ignored working-tree scratch, `git clean -fdx` will delete the progress ledger; recover from `git log` if that happens. (#1780)

### v6.0.2

- Date: 2026-06-17
- Version: v6.0.2
- Original notes: https://github.com/obra/superpowers/releases/tag/v6.0.2
- Permalink: https://whatsnew.fyi/product/superpowers/releases/v6.0.2

- **removed** — Remove the evals submodule from the shipped plugin as it broke plugin installs for some users

###### Install Fixes

- **We no longer ship the `evals` submodule.** It broke plugin installs for some users, so the eval harness now lives in its own repo, separate from the published plugin. (#1778, #1774)

### v6.0.0

- Date: 2026-06-16
- Version: v6.0.0
- Original notes: https://github.com/obra/superpowers/releases/tag/v6.0.0
- Permalink: https://whatsnew.fyi/product/superpowers/releases/v6.0.0

- **changed** — Rewrite of subagent-driven-development task review system to use a single reviewer per task instead of two, reducing token usage by approximately 50% and improving speed
- **changed** — Replace separate spec-reviewer-prompt.md and code-quality-reviewer-prompt.md files with a single task-reviewer-prompt.md file
- **changed** — Move worktrees from global ~/.config/superpowers/worktrees/ directory to project-local .worktrees/ directory
- **added** — Add support for Kimi Code harness with plugin manifest and marketplace installation
- **added** — Add support for Pi harness as a session-start extension with native skills
- **added** — Add support for Antigravity harness with direct plugin installation and bootstrap from first message
- **added** — Implement pre-flight plan validation to check for internal conflicts before the first task
- **added** — Add task-brief and review-package scripts to write task text and review diffs as files instead of pasted text
- **added** — Add Global Constraints block to plans to list rules that bind every task
- **added** — Add per-task Interfaces block to plans to specify what each task consumes and produces
- **added** — Implement single whole-branch review at end of run on the most capable model instead of re-reviewing task by task
- **changed** — Make reviewer prompts read-only and prevent reviewers from modifying the working tree or branch
- **changed** — Require explicit model specification for every dispatch instead of allowing unnamed models to inherit the session's most expensive model
- **changed** — Ban controller ability to suppress findings or pre-rate severity in reviewer judgments
- **changed** — Add file and line evidence backing to reviewer answers and progress ledger for resuming interrupted work
- **changed** — Rewrite skills' tool calls to be significantly more vendor-neutral
- **changed** — Improve brainstorming visual companion security model with authentication
- **changed** — Improve brainstorming visual companion to survive restarts and dropped connections

##### v6.0.0 (2026-06-16)

Superpowers 6.0 is a big release. The headline is a rewrite of how `subagent-driven-development` reviews each task — cheaper, stricter, and harder to game. 

While these numbers won't hold on every harness and for every workload, in our evals, Claude Code and Codex produce similar high-quality results roughly twice as fast and while spending almost 50% fewer tokens.

It also adds three new harnesses (Kimi Code, Pi, and Antigravity), gives the brainstorming visual companion a better security model, and rewrites a number of skills' tool calls to be significantly more vendor-neutral.

###### Visible Changes

- **The two per-task reviewer prompts became one.** `spec-reviewer-prompt.md` and `code-quality-reviewer-prompt.md` are gone, replaced by a single `task-reviewer-prompt.md`. If you dispatch the old files directly, switch to the new one.
- **The legacy global worktree directory is gone.** `using-git-worktrees` and `finishing-a-development-branch` no longer use `~/.config/superpowers/worktrees/`. Worktrees now land in the project — an existing `.worktrees/` or `worktrees/` if you have one, otherwise a fresh `.worktrees/` — unless you say otherwise.

###### New Harness Support

Superpowers now runs on three more harnesses. Each ships its own bootstrap, a tool-mapping reference, and tests, and each gets its own install section in the README.

- **Kimi Code** — a plugin manifest, install docs, and manifest tests; install from Kimi's marketplace or straight from the repo. (initial manifest by @qer)
- **Pi** — a session-start extension that registers the skills and injects the `using-superpowers` bootstrap. Pi has native skills, so it needs no compatibility shim.
- **Antigravity (`agy`)** — installs the plugin directly and bootstraps from the first message; verified end-to-end against the standard "make a react todo list" acceptance test.

###### Subagent-Driven Development

A long run of cost-and-quality experiments on real projects reshaped how the controller reviews each task. The old flow ran two reviewers per task and leaned on the controller's judgment for model choice and severity, and both turned out to be expensive and easy to game. The new flow runs one reviewer per task, hands work off as files instead of pasted text, and takes several judgment calls away from the controller.

- **One reviewer per task, two verdicts.** A single `task-reviewer-prompt.md` reads the task's diff once and returns both a spec-compliance verdict and a quality verdict, so one fix pass clears both. A new "can't verify from the diff" verdict flags requirements that live in untouched code, for the controller to check itself. (#1538, #1543)
- **One broad review at the end.** The run finishes with a single whole-branch review on the most capable model, instead of re-reviewing everything task by task.
- **Plans get a pre-flight read.** Before the first task, the controller checks the plan for internal conflicts — and for anything the plan asks for that a reviewer would flag as a defect — and raises it all at once, rather than stumbling into it mid-run.
- **Diffs and task text move as files.** A pasted diff parks itself permanently in the most expensive context, and a reviewer without one rebuilds it by hand — the single biggest reviewer cost. Two new scripts, `task-brief` and `review-package`, write the task text and the review diff to files for the subagent to read.
- **Every dispatch states its model.** Left to choose, controllers stopped naming a model at all — and an unnamed model quietly inherits the session's most expensive one, so one run put all 26 of its reviewers on the top tier. The templates now require a model, with guidance that reaches for cheaper tiers when the work allows.
- **The controller can't tell a reviewer what to ignore.** Real runs caught controllers coaching reviewers to skip a finding or call it "Minor at most," and the flaw shipped. Suppressing findings and pre-rating severity are now banned outri

_[Truncated at 4000 characters — full notes: https://github.com/obra/superpowers/releases/tag/v6.0.0]_