Plans
Plans are where Sinaris turns intent into project memory.
They keep the parts that should not disappear with a chat session: the project goal, milestone roadmap, phase designs, proposal decisions, and the task boundaries agents must follow.
Why It Matters
AI agents are good at local implementation, but they drift when the project direction is only stored in conversation history.
Plans give the project a durable execution contract:
- Guide captures the charter and milestone roadmap for a new project.
- Proposal captures a concrete new demand in an existing project.
- Design turns a milestone or accepted proposal into phase-level work.
- Tasks inherit context, acceptance criteria, and sequencing from the plan.
The result is not "more documentation." It is the source of truth that makes later automation safe.
What You See In The Hub
The Plan view helps you inspect the current project shape:
- active milestones and phases;
- phase status and progress;
- related design documents;
- proposal impact that should shape implementation;
- which tasks are ready, blocked, or already completed.
The prototype keeps roadmap and detail side by side so you can understand both the big picture and the current execution slice.
How Agents Use It
Agents do not start from a blank prompt. When Sinaris assigns or resumes work, it provides structured context from the relevant plan resources so the agent knows:
- what problem this phase is solving;
- which acceptance criteria matter;
- what dependencies must be respected;
- what risks or decisions are already known.
That is why Plan is upstream of Apply. Good execution depends on a stable plan.
Related Entry Points
| Need | Entry point |
|---|---|
| Start a new project | /sx-guide |
| Capture a new requirement | /sx-proposal |
| Turn direction into executable work | /sx-design |
| Inspect project state | sinaris hub |