| name | new-project-intake |
| description | When evaluating a brand-new project idea before plan-project, doing idea sparring, researching existing solutions, making a go/no-go call, bootstrapping a new project repo/workflow, and handing off to plan-project. |
| metadata | {"author":"yoskeoka","version":"1.0.0"} |
New Project Intake (Pre-Step to Workflow Step 1)
Position in workflow: This is a pre-step before plan-project (Step 1). Use this skill when the user has a raw idea but project goals/requirements are not yet ready.
What to Do
Run a structured intake flow:
-
Idea sparring
- Help the user verbalize:
- unresolved pain points
- desired user experience
- target user and usage context
- Keep this lightweight; do not force detailed requirements yet.
-
Existing-solution research
- Investigate current alternatives (products, games, tools, communities).
- Summarize what experience already exists and what is missing.
-
Checkpoint handoff note
- For any non-trivial research checkpoint, create or update
docs/issues/<descriptive-name>.md before asking the user whether to continue, pivot, compare another option, bootstrap, or stop.
- Use that note as the durable next-session handoff so web-search results, discarded comparisons, and transient technical-detail context can be dropped from chat context safely.
- Keep the note compressed and decision-oriented.
- Include:
- date
- idea / topic name
- current pain point or problem framing
- checkpoint conclusion
- live options still worth considering
- rejected/deprioritized options only when the rejection matters later
- key technical facts worth preserving
- source links for every external reference
- next questions / next candidate actions
- Update the existing note when the same line of inquiry continues. Create a separate note only when the work clearly splits into independent tracks.
-
Value test (Go/No-Go)
- Judge whether the idea has build value using:
- novelty/differentiation
- personal motivation (hobby/learning value)
- user-facing payoff
- Make the decision explicit:
- GO: proceed to bootstrap.
- NO-GO: log and stop.
-
NO-GO path
- Ensure the latest checkpoint is captured in
docs/issues/<descriptive-name>.md before stopping.
- Append summary to
docs/design-decisions/rejected-ideas.md with:
- date
- idea name
- short summary
- why no-go now
- conditions to revisit
- references
- End with a concise recommendation (pause, pivot, or re-scope).
-
GO path: project bootstrap
- Create GitHub repo if missing.
- Ensure local project directory exists under this workspace.
- Initialize docs/workflow scaffold (use
setup-workspace.sh or template structure).
- Update meta-repo management entries:
setup.sh (REPOS)
.gitignore
README.md Managed Projects
docs/project-plan.md Managed Projects
AGENTS.md workspace structure
.github/workflows/sync-workflow-to-child-repos.yml when the new repo should receive workflow sync PRs
-
Handoff to Step 1
- After bootstrap, explicitly ask:
- "Move to the new project directory and continue with
plan-project now?"
Rules
- Do not skip research and go/no-go framing when the idea is still vague.
- Keep pre-step output concise and decision-oriented.
- Do not produce a full project plan in this step; hand off to
plan-project.
- If user asks only for bootstrap work (already decided GO), skip straight to Step 5.
- Do not ask the user a checkpoint "what next?" question after substantial research until the compressed
docs/issues/ handoff note is updated.
- Keep durable notes in English unless the user explicitly asks for a different artifact language.
Deliverables
- GO case:
docs/issues/<descriptive-name>.md checkpoint handoff note with preserved research context and source links
- New repo and local project skeleton ready
- Meta-repo references updated
- Proposal to continue with
plan-project
- NO-GO case:
docs/issues/<descriptive-name>.md checkpoint handoff note with preserved research context and source links
- Entry appended to
docs/design-decisions/rejected-ideas.md
- Research-backed summary of why it is not worth starting now