Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/jsagir/mindrian-os-plugin/llms.txt

Use this file to discover all available pages before exploring further.

When you have a wicked question and need a structured argument rather than one opinion, /mos:bono spawns a governed research swarm across the subdomain × hat grid — each persona bound to its hat’s scrutiny discipline and its own wired sources — then debates the readings as a sequential inter-hat argument over a graph-proposed what-if hypothesis. The navigator confirms the topic, the hypothesis, and the ruling at three decision gates; the MECE-Minto synthesis files only on APPROVE and closes the loop into the room’s logical graph. This is not a brainstorm; it is a structured governed debate where every claim must be supported by evidence the hat was wired to at research time.

​
Usage

/mos:bono
No arguments. Larry opens with a Shape F.1 selector to confirm the topic, scope, and debate purpose before any research fires. The three decision gates — topic-confirm, hypothesis-confirm, and ruling — are the cost brakes: the research fan does not launch until the grid is scoped, and the debate does not begin until the what-if hypothesis is anchored.
/mos:bono is not autonomous_safe. It halts at three navigator decision gates and will not proceed without your confirmation at each one. No web research fires until after topic-confirm.

​
The six hats

Each hat is a distinct scrutiny discipline — not a personality, a lens:
HatFocusDiscipline
WhiteFacts & dataCite-or-retract — only verifiable claims, all sourced
RedEmotions & intuitionInstincts and gut reads — no justification required
BlackCaution & criticalACH disconfirming-first — argues what could go wrong
YellowOptimism & benefitsEvidence-backed value — argues what could go right
GreenCreativity & alternativesProvocation-marked — generates alternatives and inversions
BlueProcess & metaAnti-convergence judge — ensures the debate stays rigorous
Black hat is always re-derived, never reused from a prior run. This is a hard anti-ossification guard — the critical perspective must be fresh for every debate.

​
What happens

1

Topic-confirm gate (Gate 1)

Larry renders a Shape F.1 selector gathering: the scope (domain / subdomain grid the swarm fans over), the purpose (decision, thesis stress-test, or reframing), the substrates (which evidence sources the cells may reach), and a hypothesis seed Larry surfaces from the room graph. Before rendering, Larry reads the active JTBD via lib/hmi/jtbd-state.cjs — a validate-idea job sharpens the grid toward the risky assumption; an explore job widens it. The navigator picks; the selection becomes a typed edge. No research fires until this gate clears.
2

Domain decomposition

Larry runs Engine 1 decomposition over the confirmed scope to build the subdomain grid. Each subdomain becomes one axis of the subdomain × hat cell grid the parallel fan-out runs over.
3

Governed team assembly

Larry assembles the persona roster using the Wave-2 library-first assembleTeam: confirmed SyntheticExpert nodes first, generated gaps second, with three anti-ossification guards — a mandatory fresh slot, Black always re-derived, and a reuse cap of K < N. Each persona slot then binds to HAT_GOVERNANCE[hat]’s scrutiny discipline. assertHeterogeneity runs over the assembled cells — if two cells share an identical lens, the duplicate slot is re-drawn before any research fires.
4

Per-persona parallel research fan

The Wave-4 parallel runCellFanout launches over the subdomain × hat grid. Each cell runs extractContext → runSourceLens → wireAccept. A persona may not assert beyond its own wired source set. The part8-egress-guard is the fail-closed pre-egress gate: only the generic subdomain handle crosses toward the public web — LOCAL venture content never leaves the room.
5

Hypothesis-confirm gate (Gate 2)

The flow halts so you can confirm or edit the graph-proposed what-if hypothesis. This is the anchor the debate argues over. The confirmed hypothesis becomes the hypothesisId all per-hat derived relationships bear on.
6

Governed sequential debate

The collected cell array seeds the Wave-5 runDebate. Each hat argues its stance over its slice of the collected cells, governed by enforceGovernance and validateCitations — no hat may argue beyond its wired sources. The debate is sequential and never re-runs the research fan.
7

Ruling gate (Gate 3)

Larry halts and emits a ruling verb: supported, rejected, refined, or undecided, plus a residual-tension record naming what the debate could not close.
8

MECE-Minto synthesis

Larry applies Pyramid + MECE discipline to emit the conclusion as a governing_thought (≤ 250 chars) with 3–5 key claims. The /mos:map-unknowns Rumsfeld matrix produces the unknowns base — the open questions the debate could not close.
9

Close the loop and file

One writeCloseLoop call carries claims, relations, the conclusion, knowns, unknowns, and opportunities. The artifact files to room/solution-design/. Every node is born review_status: proposed — the engines propose, the navigator ratifies.

​
Output / Artifacts

The command produces a governed synthesis document filed to solution design: 
room/solution-design/<topic-hash>/<timestamp>.md
The artifact contains:
  • Per-hat research findings — each hat’s sourced claims from the research fan
  • Inter-hat debate record — the sequential argument chain with governance notes
  • Ruling verdict — supported | rejected | refined | undecided with residual tensions
  • MECE-Minto synthesis — governing thought + 3–5 key claims
  • Unknowns matrix — open questions from the debate, from the Rumsfeld base
  • Opportunity nodes — proposed opportunity entries for the bank

​
Pipeline position

In the PWS methodology chain, /mos:bono takes a scoped topic and returns a governed synthesis: 
/mos:map-unknowns unknowns → /mos:bono (what-if hypotheses as debate axes)
/mos:analyze-needs jobs    → /mos:bono (customer jobs as debate scope)
/mos:bono synthesis        → /mos:jtbd (active job signal updated by ruling)
The Engine 1 domain decomposition inside /mos:bono builds the subdomain grid automatically from the confirmed scope at Gate 1 — no prior command needs to have produced sub-domains first.

​
Example

/mos:bono
Larry opens Gate 1. You confirm: scope is enterprise procurement cycles in B2B SaaS, purpose is stress-test the hypothesis that 60-day procurement timelines are a myth for sub-$50K deals. Larry seeds the hypothesis from a room entry. After the research fan completes across four sub-domains × six hats, Larry halts at Gate 2. You confirm the what-if: “What if procurement timelines compress to under 30 days when the champion has budget authority?” The debate runs. Black hat finds two disconfirming studies. White hat cites three confirming deal-log analyses. At Gate 3, Larry rules: refined — the hypothesis holds for champion-led deals but fails for committee-reviewed ones. Filed synthesis: governing thought + 4 key claims + 3 open unknowns routed to the unknowns matrix.

/mos:map-unknowns

Surface the confident-but-thin assumptions before scoping the debate grid.

/mos:jtbd

Run JTBD after Bono — hat insights become context per persona.

/mos:analyze-needs

Ground the debate topic in real customer jobs before the research fan fires.

/mos:beautiful-question

Reframe the wicked question before handing it to the debate engine.

Build docs developers (and LLMs) love

Get started for freeTalk to us