| name | mission-overview-readmodel |
| description | Use when rendering, auditing, or summarizing the mission overview surface or generated mission card from canonical mission operating context. |
Mission Overview Readmodel
Goal
Show the user the smallest useful mission overview from canonical readback,
without turning Mission Overview into a transcript browser, proof console, or
control panel.
The user-facing relief is:
I can see what this mission is, what is running, what is blocked, and what
authority it has without reading the whole agent transcript.
Use This Skill When
- The agent needs to render or inspect Mission Overview.
- Chat wants to show a compact mission card through generated UI.
- A user asks for mission status, progress, governance, or task details.
- Mission Control must prove it reads the same product truth as Chat.
- A completed mission exposes a review entrypoint that routes to Mission
Overview before launching the review chat.
Do Not Use This Skill When
- The user is defining a mission; use
mission-definition-conveyor.
- The worker is executing a task; use
mission-task-operation.
- The mission is complete and the user is making a review decision; use
mission-review-loop.
Required Inputs
- Mission ref.
<mission_operating_context_tool> readback.
<current_state_projection>, proof/readiness, review surface, and runtime-proof
query refs where applicable.
- Governance policy/grant readback.
- Transcript/session refs for read-only display only.
Surface Contract
Mission Overview should contain only:
- Mission metadata needed to orient the user.
- Mission overview markdown rendered document.
- Task list; each task opens a modal with its task markdown/details.
- Progress log from the mission orchestrator with lifecycle statuses:
discovery, implementation, review and verification loop, receipts.
- Read-only live transcript window for the mission orchestrator agent.
- Governance section listing policies/grants that apply to the mission agent.
- Context-specific entrypoints such as
Review when the mission is reviewable.
Everything else should be behind details or omitted. Do not add extra default
sections just because the data is available.
Truth Rules
- The overview reads from
<mission_operating_context_tool> and related named
projections.
- The transcript is context, not control. Users do not send messages to the
mission agent from this surface.
- Human instructions to the mission must go through the chat agent, which can
translate feedback into mission tools and context updates.
- Governance display is read-only mission authority readback, not the Settings
policy editor.
- Generated UI cards are carriers. They must reconcile against canonical
mission readback before presenting approval, running, proof, or review state.
- A generated mission card may appear in Chat when the agent uses generated UI
with a mission component and mission id from mission read tools. The card is a
compact carrier of Mission Overview truth, not a separate product surface.
- Proof/readiness and blockers can be used to decide state and entrypoints, but
should not expand the default overview beyond the surface contract.
Output
Return or render:
- overview document;
- task rows with modal-ready detail refs;
- progress lifecycle rows;
- transcript read-only stream ref/status;
- governance policy/grant rows;
- review entrypoint readiness when applicable;
- source freshness and
HOLDs.
Stop Condition
Do not claim mission state when:
<mission_operating_context_tool> is missing, stale, or not for the requested
mission;
- active run, grant, task, proof, or review state is inferred from transcript;
- blockers and waits cannot be distinguished;
- UI media exists without query/readback receipts.
Receipt
Receipt must include:
- mission ref and operating-context query receipt;
- source freshness;
- UI proof if the surface was rendered;
- proof surfaces satisfied and
HOLDs;
- telemetry/canonical applicability note for read-only overview actions.