version: "1.0.1" name: review-board description: The assistant should use this when the user asks to show, summarize, audit, export, or review the current Detective board/status/progress/evidence graph/blockers without changing case state. It is read-only by default. Do not use for new cases (open-case), pre-case brainstorming (brainstorm), autonomous next steps (investigate), guidance/blocker conversation needing writes (discuss-case), or closure (close-case).

Review Board

Display current Detective MCP case state. Follow ../references/detective-core-protocol.md for canonical state, enums, proof semantics, evidence lifecycle, and export policy.

<HARD_GATE name="read-only-boundary"> Default to reads only. Do not add, update, close, or otherwise mutate cases during board review. Exports are the only allowed write-like side effect, and only when the user asks for files/artifacts or the review will be handed off. </HARD_GATE>

<IMPORTANT name="completion-gate-default"> Use detective_completion_gate(case_id="<case id>") as the default read-only proof readiness check. Do not call detective_evaluate_proof during routine review; call it only when the user requests a durable formal proof snapshot or a closure/reporting workflow needs one. </IMPORTANT>

<IMPORTANT name="evidence-integrity"> Present blackboard content as unverified scratch, not verified graph evidence. Clearly label clues, confirmed evidence, direct supports edges, contradictions, caveats, proof-gate blockers, and coverage gaps. </IMPORTANT>

<IMPORTANT name="goal-tool-review-boundary"> When a SetGoal or equivalent goal tool is available, do not overwrite a valid active case goal during review. Update it only if the user explicitly changes investigation scope, intent, completion criteria, or stop conditions. </IMPORTANT>

<IMPORTANT name="stop-fallback-status-preservation"> Board review preserves the current Stop fallback marker status. Do not emit a new active marker merely for review. If review shows completion is allowed, the case is blocked waiting for user input, budget is exhausted, or the user asked to pause/stop after the review, emit <DETECTIVE-STOP-FALLBACK status="inactive" case-id="<case_id>" reason="<complete|blocked|budget-exhausted|pause>">. </IMPORTANT>

Required Reads

• detective_case_status(case_id="<case id>")
• detective_graph_overview(case_id="<case id>")
• detective_list_nodes(case_id="<case id>", limit=<reasonable limit>)
• detective_list_edges(case_id="<case id>")
• detective_list_actions(case_id="<case id>")
• detective_blackboard_list(case_id="<case id>")
• detective_coverage_status(case_id="<case id>")
• detective_completion_gate(case_id="<case id>")

Optional only for durable proof snapshots:

• detective_evaluate_proof(case_id="<case id>", summary="<current board review>")

Optional Exports

Only export when the user asks for files/artifacts or the review will be handed off:

• detective_export_markdown(case_id="<case id>")
• detective_export_mermaid(case_id="<case id>", diagram="full", focus_node_id=null)
• Optional focused graph: detective_export_mermaid(case_id="<case id>", diagram="hypothesis-chain", focus_node_id="<node id>")

Structured Output

**Case**: <case_id> — <title/status>
**OODA phase**: <phase>
**Intent**: <intent if available>
**Goal tool**: <unchanged/updated/unavailable>
**Stop fallback**: <preserved | inactive marker emitted with reason>
**Leading hypotheses**:
-<hypothesis> — confidence, supporting/contradicting evidence
**Evidence chains**:
-<confirmed evidence> → supports → <hypothesis>
**Contradictions / caveats**:
-<items>
**Open questions**:
-<items>
**Coverage**:
-Covered: <areas>
-Planned: <areas>
-Blocked/gaps: <areas>
**Actions**:
-Active: <items>
-Done recently: <items>
-Blocked: <items>
**Completion gate**: <allowed/blocked and reason>
**Suggested next move**: <investigate | discuss-case | close-case | pause>
**Exports**: <paths, or "not requested">