project-seed
Skill0 starsChanged 4 months agoMITView on GitHub
Use when scoping, expanding or stress-testing an idea. Use with /project-seed.
In this repository42 other files it gives its agents.
- .claude/CLAUDE.mdCLAUDE.md
- content-rewriteSkill
- content-writeSkill
- core-auditSkill
- core-bootstrapSkill
- core-commitSkill
- core-explainSkill
- core-finalizeSkill
- core-orchestrateSkill
- core-pullSkill
- core-setupSkill
- core-updateSkill
- design-convertSkill
- design-shipSkill
- design-tokensSkill
- dev-inspectSkill
- dev-learnSkill
- dev-manualSkill
- dev-optimizeSkill
- dev-securitySkill
- dev-shipSkill
- dev-tweakSkill
- game-debugSkill
- game-optimizeSkill
- game-shipSkill
- game-tweakSkill
- marketing-contentSkill
- marketing-researchSkill
- marketing-screenshotsSkill
- project-addSkill
- project-appSkill
- project-memorySkill
- project-planSkill
- project-removeSkill
- project-researchSkill
- project-syncSkill
- project-todoSkill
- project-tunnelSkill
- team-issuesSkill
- team-outsourceSkill
- team-reviewSkill
- team-verifySkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 67
Every mode ends via shared/THINKING-OUTPUT.md.
- Very specific: line 75
NEVER use blockquote syntax (>) for displaying content — causes unreadable white background in dark terminals
- Distinctive: line 76
NEVER use inline code backticks for emphasis on regular words — use bold or plain text
- Distinctive: line 81
Follow the Language Policy in CLAUDE.md.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Seed
Transform any idea, concept, feature, or task assignment into a structured seed document — and optionally expand (brainstorm) or stress-test (critique) it. Three modes share one output pipeline into /project-plan:
- seed (default): scope an idea, design, or task into a seed document through targeted questions and synthesis; also syncs an existing seed with project state
- brainstorm: creatively expand an idea via brainstorm techniques, one technique at a time through dialogue
- critique: stress-test an idea via multi-perspective analysis techniques
When to Use
- User starts with
/project-seed(with or without arguments) — mode dispatch below decides the route - New or vague idea, design, or task needs scoping → seed mode
- User wants variations, alternatives, or creative expansion → brainstorm mode
- User wants weaknesses identified, assumptions tested, failure modes found → critique mode
PHASE 0: Mode Dispatch
Keyword routing — first whitespace-delimited token of the invocation args, case-insensitive:
brainstorm→ brainstorm mode; remaining args = topic inputcritique→ critique mode; remaining args = topic input- anything else (including no args) → seed mode; full args = topic input (on an existing seed the
sync,critiqueandbrainstormroutes are offered inside seed-mode intake)
This parser covers both slash invocation (
/project-seed brainstorm dark-mode toggle) and Skill-tool invocation (args: "critique <topic>").- Intent routing (no mode keyword, natural-language trigger):
- weaknesses / assumptions / risks / "poke holes" / "stress-test" → critique
- variations / alternatives / "wild ideas" / expand → brainstorm
- new idea / scope / task intake, or no clear signal → seed (default)
- Two modes equally signalled → one AskUserQuestion (header
"Mode", options Seed / Brainstorm / Critique) — never guess between brainstorm and critique.
- Dispatch:
Todo: Read
.claude/skills/project-seed/references/mode-{seed|brainstorm|critique}.md— follow it to completion, passing along the parsed topic input (if any).
Continue Step
Every mode ends via shared/THINKING-OUTPUT.md; for concept-scope output that file closes with its § Continue step (chain another mode in this session, hand off to /project-plan, or stop). The dispatcher itself does nothing after dispatch.
Best Practices
Terminal Formatting
- NEVER use blockquote syntax (
>) for displaying content — causes unreadable white background in dark terminals - NEVER use inline code backticks for emphasis on regular words — use bold or plain text
- Backticks only for actual code, file paths, and command references
Language
Follow the Language Policy in CLAUDE.md.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

