From ideation
Generate a /goal command that runs an ideation project to completion UNATTENDED by driving /ideation:autopilot. The /goal is a durability wrapper — it keeps autopilot going across hours/sessions and recovers from failures, while autopilot's Workflow engine does the dependency-ordered dispatch. Reads the contract, builds the /goal, copies it to the clipboard, and prints it. Use when the user says 'goal', 'run as goal', 'goal prompt', or wants to run the project long-haul/unattended rather than watching /ideation:autopilot interactively.
How this skill is triggered — by the user, by Claude, or both
Slash command
/ideation:get-goal-prompt [path/to/contract.md][path/to/contract.md]This skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Build a `/goal` command that runs an ideation project to completion **unattended** by
Build a /goal command that runs an ideation project to completion unattended by
driving /ideation:autopilot, then copy it to the clipboard.
The layered model: /ideation:autopilot is the deterministic Workflow engine that
dispatches phases in dependency order. A /goal is an optional durability wrapper
around it — it keeps working across turns/hours/sessions and can recover from a failed
phase (fix the spec, re-run) in ways the engine can't. So this /goal should drive
autopilot, not re-describe per-phase execution. One engine underneath; /goal just
makes the run durable and self-healing.
Reach for this over plain /ideation:autopilot when the user wants to start the project
and walk away (long-haul, multi-session). For a run they'll watch, plain
/ideation:autopilot is lighter.
$ARGUMENTS, use it../docs/ideation/*/contract.md and pick the most recently modified.Read contract.md (and the sibling contract-data.json if present) and extract:
git log --oneline --grep="<slug>"
(match the slug-qualified spec path, not the bare spec-phase-N.md filename, which
collides across projects). This is only to tell the user where the run will resume
from; autopilot itself re-derives this on each run.If every phase is already committed, tell the user the project is complete and stop.
The /goal drives autopilot to completion. Keep it short — autopilot + the specs hold
all the per-phase detail, so the /goal does not list per-phase steps.
/goal Drive the {project-name} ideation project to completion by running /ideation:autopilot.
1. Run /ideation:autopilot {contract-path}.
2. autopilot executes the specs in dependency order via its Workflow engine and returns a summary of completed / failed / skipped phases.
3. If any phases failed: read each failure, fix the spec or the implementation, then run /ideation:autopilot {contract-path} again — it automatically skips already-committed phases and resumes from what remains.
4. Repeat until autopilot reports every phase committed with passing validation and zero failed/skipped phases.
Done when: git log shows a commit referencing every phase's slug-qualified spec path, and a final /ideation:autopilot run reports nothing left to do. If the same phase fails 3 times without progress, stop and report which phase and why.
Fill in {project-name} and {contract-path}. If some phases are already committed,
add one line after the template: Phases {titles} are already committed — autopilot will skip them.
Copy the full /goal command to the clipboard:
echo '<goal command>' | pbcopy # macOS
On Linux, try xclip -selection clipboard or xsel --clipboard.
Print the command so the user can see what they're about to run:
Copied to clipboard. Paste to start an unattended run:
/goal Drive the {project-name} ideation project to completion ...
Remind the user:
/auto if needed)."That's it. Once the command is on the clipboard, this skill's job is done. The /goal
wrapper drives /ideation:autopilot, which drives the Workflow engine.
npx claudepluginhub nicknisi/ideation --plugin ideationCreates bite-sized, testable implementation plans from specs or requirements, with file structure and task decomposition. Activates before coding multi-step tasks.