pi-goals
Make a short list of goals in one Markdown plan file. This is easy to review, and a subagent can check whether each goal is complete.
The plan file looks like this:
## <short plan title>
<context: one short paragraph. What the human wants and why.>
### User-visible result
<one concrete sentence naming the final artifact or behavior the human will inspect>
### User voice
- │ "<the human's requirement, quoted in full word for word (with spelling fixes)>"
### Goals
1. [ ] goal: <one short judgeable imperative outcome>
- subtle failure mode: <a way this could look done but isn't>
- discriminator: <the concrete observation that tells real success from that failure>
- tasks:
1. [ ] <subtask>
- evidence: (empty until sign-off)
### Future work / out of scope
### Log
### Interview
### Learnings
### Papercuts - problems, gotchas, suggestions
Related work
Like pi-milestones and burneikis/pi-plan, it guides rather than guards. The reminder cadence is copied from tintinweb/pi-tasks and the resync-after-compaction from tmonk/pi-goal-x.
Install
pi install npm:@wassname2/pi-goals
Or for development:
git clone https://github.com/wassname/pi-goals && cd pi-goals && npm install
pi -e ./src/index.ts
Use
/goals CSV export for the report view
/goals enters plan mode and starts a conversation; the objective is an optional seed. From there:
- Plan. The agent explores read-only and drafts the plan.
- Review. After Pi settles, the full plan is printed in the transcript. Check that User-visible result names the final artifact or behavior you expect. The menu offers Ready, Refine, Edit, or Cancel. Refine collects short notes. Edit opens the full plan in Pi's editor.
- Work. Ready is the only review action that starts work. The agent ticks subtasks, appends to
## Logand## Learnings, fillsevidence:, and callsCompleteGoalwhen a discriminator is satisfied. Every human reply and Refine note in plan mode is saved verbatim under## Interview. After eight turns without a change above## Log, the working set is sent back with a short upkeep reminder.
Optional persistent perspective requires the separate pi-subagents package (pi install npm:pi-subagents, then restart or reload Pi). /goals steward on forks one non-writing Oracle
when Ready is selected. Work waits for its plan decision. The child process exits after the
review, while its session is retained. The first CompleteGoal call resumes that same session for a
trajectory and scope check; after approval, a second call runs the normal fresh evidence judge. The
steward cannot complete goals or make unresolved human decisions. /goals steward off disables it.
Between reviews there is no running child process or model call. The retained session receives a
bounded contract view: evidence detail stays with the fresh judge, approved checkbox state is
normalized, and current goal status remains visible. The Oracle's profile includes inspection-only
bash by contract; pi-goals rejects its decision if pi-subagents reports a file-mutation effect. This
is not an OS sandbox. If pi-subagents is absent, Ready stays in planning after a visible RPC timeout;
install it, retry Ready, or use /goals steward off. The integration is process-local and does not
require pi-intercom.
Other commands: /goals --clear disconnects this session from its active plan, preserving the
versioned file on disk; /goals --auto [minutes|off] continues active goals after the agent settles
and then on that interval. It pauses after two automatic wakes with no working-plan change;
/goals --judge <model-ref> picks a sign-off judge model (default: your current session model, else
pi's default); /goals steward [on|off|status] controls the optional persistent plan steward. The
older -- forms remain only for the existing clear, auto, and judge controls.
Prompts
All model-facing text lives in src/prompts.ts, in flow order.
Develop
pi -e ./src/index.ts # load locally
npm test # all unit, flow, and Pi RPC tests
npm run test:rpc # Pi RPC review flow with a local offline model
npm run typecheck
npm run lint
License
MIT