requesting-code-review had "review after each batch (3 tasks)" for
executing-plans, which leaked into subagent-driven-development as a
check-in cadence. Replaced with flexible "each task or at natural
checkpoints" and added explicit continuous execution directive to
subagent-driven-development.
- commit .codex-plugin/plugin.json and brand assets in this repo
- sync tracked Codex plugin files instead of generating or seeding them
- honor upstream gitignored files during rsync
- cover the new sync behavior with regression tests
Speak directly to AI agents at the top of CLAUDE.md: reframe slop
PRs as harmful to their human partner, give a concrete pre-submission
checklist, and explicitly authorize pushing back on vague instructions.
CLAUDE.md (symlinked to AGENTS.md) covers every major rejection
pattern from auditing the last 100 closed PRs (94% rejection rate):
AI slop, ignored PR template, duplicates, speculative fixes, domain-
specific skills, fork confusion, fabricated content, bundled changes,
and misunderstanding project philosophy.
Move bootstrap injection from experimental.chat.system.transform to
experimental.chat.messages.transform, prepending to the first user
message instead of adding a system message.
This avoids two issues:
- System messages repeated every turn inflate token usage (#750)
- Multiple system messages break Qwen and other models (#894)
Tested on OpenCode 1.3.2 with Claude Sonnet 4.5 — brainstorming skill
fires correctly on "Let's make a React to do list" prompt.
The bootstrap text advertised a configDir-based skills path that didn't
match the runtime path (resolved relative to the plugin file). Tests
used yet another hardcoded path and referenced a nonexistent lib/ dir.
- Remove misleading skills path from bootstrap text; the agent should
use the native skill tool, not read files by path
- Fix test setup to create a consistent layout matching the plugin's
../../skills resolution
- Export SUPERPOWERS_SKILLS_DIR from setup.sh so tests use a single
source of truth
- Add regression test that bootstrap doesn't advertise the old path
- Remove broken cp of nonexistent lib/ directory
Fixes#847
Move bootstrap injection from experimental.chat.system.transform to
experimental.chat.messages.transform, prepending to the first user
message instead of adding a system message.
This avoids two issues:
- System messages repeated every turn inflate token usage (#750)
- Multiple system messages break Qwen and other models (#894)
Tested on OpenCode 1.3.2 with Claude Sonnet 4.5 — brainstorming skill
fires correctly on "Let's make a React to do list" prompt.
The bootstrap text advertised a configDir-based skills path that didn't
match the runtime path (resolved relative to the plugin file). Tests
used yet another hardcoded path and referenced a nonexistent lib/ dir.
- Remove misleading skills path from bootstrap text; the agent should
use the native skill tool, not read files by path
- Fix test setup to create a consistent layout matching the plugin's
../../skills resolution
- Export SUPERPOWERS_SKILLS_DIR from setup.sh so tests use a single
source of truth
- Add regression test that bootstrap doesn't advertise the old path
- Remove broken cp of nonexistent lib/ directory
Fixes#847
Two bugs caused the brainstorm server to self-terminate within 60s:
1. ownerAlive() treated EPERM (permission denied) as "process dead".
When the owner PID belongs to a different user (Tailscale SSH,
system daemons), process.kill(pid, 0) throws EPERM — but the
process IS alive. Fixed: return e.code === 'EPERM'.
2. On WSL, the grandparent PID resolves to a short-lived subprocess
that exits before the first 60s lifecycle check. The PID is
genuinely dead (ESRCH), so the EPERM fix alone doesn't help.
Fixed: validate the owner PID at server startup — if it's already
dead, it was a bad resolution, so disable monitoring and rely on
the 30-minute idle timeout.
This also removes the Windows/MSYS2-specific OWNER_PID="" carve-out
from start-server.sh, since the server now handles invalid PIDs
generically at startup regardless of platform.
Tested on Linux (magic-kingdom) via Tailscale SSH:
- Root-owned owner PID (EPERM): server survives ✓
- Dead owner PID at startup (WSL sim): monitoring disabled, survives ✓
- Valid owner that dies: server shuts down within 60s ✓
Fixes#879
ownerAlive() treated EPERM (permission denied) the same as ESRCH
(process not found), causing the server to self-terminate within 60s
whenever the owner process ran as a different user. This affected WSL
(owner is a Windows process), Tailscale SSH, and any cross-user
scenario.
The fix: `return e.code === 'EPERM'` — if we get permission denied,
the process is alive; we just can't signal it.
Tested on Linux via Tailscale SSH with a root-owned grandparent PID:
- Server survives past the 60s lifecycle check (EPERM = alive)
- Server still shuts down when owner genuinely dies (ESRCH = dead)
Fixes#879
The session directory now contains two peers: content/ (HTML served to
the browser) and state/ (events, server-info, pid, log). Previously
all files shared a single directory, making server state and user
interaction data accessible over the /files/ HTTP route.
Also fixes stale test assertion ("Waiting for Claude" → "Waiting for
the agent").
Reported-By: 吉田仁
Metadata files (.server-info, .events, .server.pid, .server.log,
.server-stopped) were stored in the same directory served over HTTP,
making them accessible via the /files/ route. They now live in a .meta/
subdirectory that is not web-accessible.
Also fixes a stale test assertion ("Waiting for Claude" → "Waiting for
the agent").
Reported-By: 吉田仁
The subagent review loop (dispatching a fresh agent to review plans/specs)
doubled execution time (~25 min overhead) without measurably improving plan
quality. Regression testing across 5 versions (v3.6.0 through v5.0.4) with
5 trials each showed identical plan sizes, task counts, and quality scores
regardless of whether the review loop ran.
Changes:
- writing-plans: Replace subagent Plan Review Loop with inline Self-Review
checklist (spec coverage, placeholder scan, type consistency)
- writing-plans: Add explicit "No Placeholders" section listing plan failures
(TBD, vague descriptions, undefined references, "similar to Task N")
- brainstorming: Replace subagent Spec Review Loop with inline Spec Self-Review
(placeholder scan, internal consistency, scope check, ambiguity check)
- Both skills now use "look at it with fresh eyes" framing
Testing: 5 trials with the new skill show self-review catches 3-5 real bugs
per run (spawn positions, API mismatches, seed bugs, grid indexing) in ~30s
instead of ~25 min. Remaining defects are comparable to the subagent approach.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Two bugs caused the brainstorm server to self-terminate within 60s:
1. ownerAlive() treated EPERM (permission denied) as "process dead".
When the owner PID belongs to a different user (Tailscale SSH,
system daemons), process.kill(pid, 0) throws EPERM — but the
process IS alive. Fixed: return e.code === 'EPERM'.
2. On WSL, the grandparent PID resolves to a short-lived subprocess
that exits before the first 60s lifecycle check. The PID is
genuinely dead (ESRCH), so the EPERM fix alone doesn't help.
Fixed: validate the owner PID at server startup — if it's already
dead, it was a bad resolution, so disable monitoring and rely on
the 30-minute idle timeout.
This also removes the Windows/MSYS2-specific OWNER_PID="" carve-out
from start-server.sh, since the server now handles invalid PIDs
generically at startup regardless of platform.
Tested on Linux (magic-kingdom) via Tailscale SSH:
- Root-owned owner PID (EPERM): server survives ✓
- Dead owner PID at startup (WSL sim): monitoring disabled, survives ✓
- Valid owner that dies: server shuts down within 60s ✓
Fixes#879
ownerAlive() treated EPERM (permission denied) the same as ESRCH
(process not found), causing the server to self-terminate within 60s
whenever the owner process ran as a different user. This affected WSL
(owner is a Windows process), Tailscale SSH, and any cross-user
scenario.
The fix: `return e.code === 'EPERM'` — if we get permission denied,
the process is alive; we just can't signal it.
Tested on Linux via Tailscale SSH with a root-owned grandparent PID:
- Server survives past the 60s lifecycle check (EPERM = alive)
- Server still shuts down when owner genuinely dies (ESRCH = dead)
Fixes#879
The session directory now contains two peers: content/ (HTML served to
the browser) and state/ (events, server-info, pid, log). Previously
all files shared a single directory, making server state and user
interaction data accessible over the /files/ HTTP route.
Also fixes stale test assertion ("Waiting for Claude" → "Waiting for
the agent").
Reported-By: 吉田仁
Metadata files (.server-info, .events, .server.pid, .server.log,
.server-stopped) were stored in the same directory served over HTTP,
making them accessible via the /files/ route. They now live in a .meta/
subdirectory that is not web-accessible.
Also fixes a stale test assertion ("Waiting for Claude" → "Waiting for
the agent").
Reported-By: 吉田仁
The subagent review loop (dispatching a fresh agent to review plans/specs)
doubled execution time (~25 min overhead) without measurably improving plan
quality. Regression testing across 5 versions (v3.6.0 through v5.0.4) with
5 trials each showed identical plan sizes, task counts, and quality scores
regardless of whether the review loop ran.
Changes:
- writing-plans: Replace subagent Plan Review Loop with inline Self-Review
checklist (spec coverage, placeholder scan, type consistency)
- writing-plans: Add explicit "No Placeholders" section listing plan failures
(TBD, vague descriptions, undefined references, "similar to Task N")
- brainstorming: Replace subagent Spec Review Loop with inline Spec Self-Review
(placeholder scan, internal consistency, scope check, ambiguity check)
- Both skills now use "look at it with fresh eyes" framing
Testing: 5 trials with the new skill show self-review catches 3-5 real bugs
per run (spawn positions, API mismatches, seed bugs, grid indexing) in ~30s
instead of ~25 min. Remaining defects are comparable to the subagent approach.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Four templates: bug report (with environment table and platform-vs-plugin
gate), feature request (with problem statement and core-appropriateness
question), IDE/platform support request, and a config that disables
blank issues and redirects questions to Discord.
Requires contributors to articulate the problem they're solving,
confirm human review, document eval methodology, and check for
duplicate PRs. Informed by patterns in ~90 closed-without-merge PRs.
- Remove chunk-based plan review in favor of single whole-plan review
- Add Calibration sections to both reviewer prompts so only serious
issues block approval
- Reduce max review iterations from 5 to 3
- Streamline reviewer checklists (spec: 7→5, plan: 7→4 categories)
Subagent Context Isolation:
All delegation skills (brainstorming, parallel agents, code review,
subagent-driven development, writing plans) now explicitly instruct the
dispatching agent to construct review context from scratch — never
forward session history to subagents.
This fixes a problem observed with Codex, where subagents inherited the
full parent session context including the dispatcher's internal
reasoning, prior conversation, and user-facing tone. Reviewers that
inherited this context behaved as if they were the lead developer rather
than a reviewer — they'd reject reasonable code for not matching
unstated preferences, demand rewrites beyond scope, and treat advisory
feedback as blocking. The fix is simple: the dispatcher crafts precisely
what each subagent needs (the spec, the code, the review criteria) and
nothing else. This keeps reviewers focused on the work product, not the
thought process that produced it, and also preserves the dispatcher's
own context window for coordination.
Zero-Dependency Brainstorm Server:
The brainstorm visual companion server has been rewritten from scratch
as a single zero-dependency Node.js file. The previous implementation
vendored Express, ws, chokidar, and 714 npm packages (84,000+ lines of
third-party code) — a supply chain surface area that was
disproportionate to what the server actually does.
The new server.js (~340 lines) implements everything with Node built-ins
only: RFC 6455 WebSocket protocol, HTTP server with template wrapping,
fs.watch with debounce, and lifecycle management.
731 files changed, 1,700 insertions, 85,000 deletions. The entire
vendored node_modules/ directory is gone.
Server Lifecycle Management:
The brainstorm server now automatically shuts down when no longer
needed, preventing orphaned processes. Owner process tracking captures
the harness PID at startup and checks every 60 seconds. 30-minute idle
timeout as fallback. The visual companion guide now instructs agents to
check .server-info before each write and restart if .server-stopped
exists.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Subagents should never inherit the parent session's context or history.
The dispatcher constructs exactly what each subagent needs, keeping
both sides focused: the subagent on its task, the controller on
coordination.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
$PPID inside start-server.sh is the ephemeral shell the harness spawns
to run the script — it dies immediately when the script exits, causing
the server to shut down after ~60s. Now resolves grandparent PID via
`ps -o ppid= -p $PPID` to get the actual harness process (e.g. claude).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
start-server.sh passes $PPID as BRAINSTORM_OWNER_PID to the server.
The server checks every 60s if the owner process is still alive
(kill -0). If it's gone, the server shuts down immediately —
deletes .server-info, writes .server-stopped, exits cleanly.
Works across all harnesses (CC, Codex, Gemini CLI) since it
tracks the shell process that launched the script, which dies
when the harness dies.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Server tracks activity (HTTP requests, WebSocket messages, file
changes) and exits after 30 minutes of inactivity. On exit, deletes
.server-info and writes .server-stopped with reason. Visual companion
guide now instructs agents to check .server-info before each screen
push and restart if needed. Works on all harnesses, not just CC.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Delete 717 files: index.js, package.json, package-lock.json, and
the entire node_modules directory (express, ws, chokidar + deps).
Update start-server.sh to use server.js. Remove gitignore exception.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Complete zero-dep brainstorm server. Uses knownFiles set to
distinguish new screens from updates (macOS fs.watch reports
'rename' for both). All 56 tests pass (31 unit + 25 integration).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Implements RFC 6455 handshake, frame encoding/decoding for text
frames. All 31 unit tests pass.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace vendored node_modules (714 files) with a single server.js
using only Node built-ins. Spec covers WebSocket protocol, HTTP
serving, file watching, and static file serving. Tests written
before implementation (TDD).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The spec review loop (dispatch spec-document-reviewer subagent, iterate
until approved) existed in the prose "After the Design" section but was
missing from both the checklist and the process flow diagram. Since agents
follow the diagram and checklist more reliably than prose, the spec review
step was being skipped entirely.
Added step 7 (spec review loop) to the checklist and a corresponding
"Spec review loop" → "Spec review passed?" node pair to the dot graph.
Tested with claude --plugin-dir and claude-session-driver: worker now
correctly dispatches the spec-document-reviewer subagent after writing
the design doc and before presenting to the user for review.
Fixes#677.
Use escaped double quotes instead of single quotes around
${CLAUDE_PLUGIN_ROOT} path in hooks.json.
Single quotes fail on Windows (cmd.exe doesn't recognize them as path
delimiters) and on Linux when the shell doesn't expand the variable.
Verified the fix works across all combinations:
- macOS bash, path without spaces: pass
- macOS bash, path with spaces: pass
- Windows cmd.exe, path without spaces: FAILED with single quotes, PASS with double quotes
- Windows cmd.exe, path with spaces: FAILED with single quotes, PASS with double quotes
- Windows Git Bash: pass (both quote styles work here)
Testing was done on a Windows 11 (NT 10.0.26200.0) dev box with
Claude Code 2.1.72 and Git for Windows. The single-quote bug only
manifests when cmd.exe is the executing shell (no Git Bash fallback),
which explains why some users hit it and others don't.
Closes#577, closes#644.
start-server.sh now runs npm install if node_modules is missing.
Fixes broken server when superpowers is installed as a plugin (node_modules
are in .gitignore and not included in the clone).
The server now writes its startup JSON to $SCREEN_DIR/.server-info.
Agents that launch the server via background execution (where stdout is
hidden) can read this file to get the URL, port, and screen_dir.
The visual companion docs now give concrete launch commands per platform:
Claude Code (default mode), Codex (auto-foreground via CODEX_CI), Gemini CLI
(--foreground with is_background), and a fallback for other environments.
Moves lib/brainstorm-server/ → skills/brainstorming/scripts/ so the
brainstorming skill uses relative paths (scripts/start-server.sh) instead
of ${CLAUDE_PLUGIN_ROOT}/lib/brainstorm-server/. This follows the
agentskills.io specification for portable, cross-platform skills.
Updates visual-companion.md references and test paths. All tests pass.
The tool mapping table is now @referenced directly in GEMINI.md so Gemini
CLI always has it in context when processing skills, rather than requiring
Gemini to find and read a reference file from within the skill.
Maps all Claude Code tool names to Gemini CLI equivalents (read_file,
write_file, replace, run_shell_command, grep_search, glob, write_todos,
activate_skill, etc.). Notes that Gemini CLI has no subagent support.
Updates using-superpowers to reference GEMINI.md in instruction priority
and link to the new gemini-tools.md reference alongside codex-tools.md.
Symlinks inside .gemini/ don't work on Windows. Moving
gemini-extension.json and GEMINI.md to the repo root means
the extension root IS the repo root, so skills/ is found
naturally without symlinks.
Symlinks .gemini/skills -> ../skills so the extension bundles
all Superpowers skills. Without this, skills are only found when
running from the repo workspace, not when installed as an extension.
Remove the custom MCP server (find_skills/use_skill tools) and
force-invoke GEMINI.md. Gemini CLI natively supports the Agent Skills
format — our skills/ directory works as-is.
GEMINI.md now uses @import to inline using-superpowers content at
session start. Needs testing to verify @import resolves relative
to the extension root.
Add experimental Gemini CLI extension with MCP server that exposes
skills as individual tools. Infrastructure works but auto-triggering
skills is blocked by Gemini CLI treating context files as advisory
rather than executable instructions.
See issue #128 for detailed findings.
- gemini-extension.json manifest
- MCP server with individual skill tools
- GEMINI.md bootstrap attempts (don't work)
- Installation documentation
