# Runtime Port (PROTOCOL.md §9.1), emitted by Phase 5 (Stability) at # evidence_level `signaled`. Declares every callable workflow of the Bentonville # studio's operating design as a Skill. # # NOTE ON PLAYBOOKS: reconciled v1.1 against the 11 playbooks /fw-flow actually # emitted (branch fw-flow). Paths resolve on merge, not on this branch. Flow's 03, # 05, 09 and 11 have no skill here — they are venture-lifecycle and operating-week # workflows whose decision rights sit with Tulsa, not operating skills of this seat. # # NOTE ON SCOPE: these are operating skills for a studio that does not yet exist. # They describe how the seat would be run, not automation to be provisioned. This # run produces documents, not software (PROJECT-CONTEXT, "Tooling"). # AMENDMENT v1.2 (Phase 7, /fw-entropy) — three declared skills added, nothing # else changed. Provenance stated rather than edited in quietly: these are # Stability's to own; this amendment closes EntropyAudit E-01, E-02, E-03 and # ConsonanceCheck C-8, which are the writers behind surfaces that already render. # # DELIBERATELY NOT RESOLVED HERE, because they are open criticals or another # skill's to close, and pre-empting them would bake in one side: # G-6 / C-7 — this file says Tier 1/T/E; playbooks 04, 11, 12 say Tier 1/2/3. # 38 occurrences across 7 locations. /fw-consonance's. The three # skills added below therefore name the RETAINED RIGHT rather than # a tier, introducing no new occurrence of either vocabulary. # G-7 — `venture-leader-naming` below still reads "carries the second # profits-interest grant", which contradicts decision 007's "no # new equity attaches". Flow resolved it (vest a tranche of the # day-one grant) and the reference render already renders # "tranche vested". Text left unrepointed: /fw-consonance's C-3, # which found the same contradiction in 7 places, not 1. # C-1 — two incompatible readiness bar structures, both locked. No skill # below asserts a bar item count. # C-2 — the inverted "Reading A" and the four residency calendars. No # skill below states a residency month. # C-6 — the Tulsa answer window is 5 days here and 10 in playbook 04. # No skill below states an SLA. artifact: SkillManifest phase: stability version: 1.2 status: working-draft evidence_level: signaled track: agency-studio supersedes: 1.1 skills: - id: cohort-grant-calendar name: Cohort grant calendar description: > Hold the QSBS timing rule. A cohort's profits interests are granted, papered and filed on one date, before any venture that cohort touches is incorporated (25:23). Where the grant calendar and a formation date collide, the incorporation waits or the grant moves earlier — never the reverse. domain: resident-economics trigger: scheduled schedule: per cohort start, and reviewed whenever a venture formation date is proposed ownership: human playbook: define/playbooks/07-resident-economics-and-status-ladder.md inputs: [cohort start date, proposed venture formation dates] outputs: [papered grant date, resident term sheet, QSBS eligibility line per resident] depends_on_capabilities_note: resident-term-sheet is emitted at define/resident-term-sheet.md depends_on_capabilities: [resident-economic-record, resident-term-sheet] reads_memory: [define/, define/playbooks/, define/records/decisions/] writes_memory: [define/records/residents/] escalation: > Tier 2. A formation date that cannot move is a Tulsa decision with the QSBS reason attached (stability-specs §2.7). tags: [core, qsbs, hard-rule] - id: readiness-review name: Readiness review description: > Monthly update of each resident's record against decision 006's six bar items, evidenced by artifacts rather than assessment scores. Items 4 and 5 take a track-specific form per Entrepreneur / Engineer in Residence (stability-specs §3.2a). Dated reviews at month 6 and month 12 decide reassign or exit. domain: readiness-bar trigger: scheduled schedule: monthly, with dated gates at month 6 and month 12 of each residency ownership: hybrid playbook: define/playbooks/06-resident-intake-and-readiness-bar.md inputs: [resident record, artifact register entries] outputs: [updated resident record, trajectory call, month-6 and month-12 decisions] depends_on_capabilities: [readiness-bar, resident-record, track-differentiated-bar] reads_memory: [define/, define/intake/, define/playbooks/, define/registry/artifacts/, define/records/residents/] writes_memory: [define/records/residents/] escalation: > Shared with the talent function by construction — the record is the mechanism that stops fills and venture-leaders diverging silently. tags: [core, stage-gate, joint-with-talent] - id: venture-leader-naming name: Venture leader naming description: > Recommend a resident as venture leader on a virtual portfolio company (47:01–47:06) once the stated subset of decision 006's six items carries artifact evidence — ALL THREE CORE items plus at least one of the three track items (four of six), assessed on the pair, not only on the individual (015). Repointed at decision 006 Amendment 2, which closed /fw-consonance C-1: the bar existed in two incompatible structures, both locked. The prior text ("items 1, 3 and 6 plus at least one of 2 or 5") was the same strength against the superseded six-common-item form. Bentonville recommends the person; Tulsa concurs on the venture. Carries **a vest of a tranche of the day-one grant — never a second grant**, so it is a status event and an economic one at the same time. Repointed at C-3/G-7: this text said "carries the second profits-interest grant", which contradicts decision 007's "No new equity attaches" and would re-introduce the per-venture incentive 40:44 removed. Flow resolved it; the render already showed "tranche 2 vested — no new grant issued". The manifest was the last holdout. domain: venture-leader-naming trigger: event event: a resident reaches the stated subset of bar evidence and a virtual portfolio company is available ownership: human playbook: define/playbooks/07-resident-economics-and-status-ladder.md inputs: [resident record, virtual portfolio company status] outputs: [naming recommendation, Tulsa concurrence, tranche vest recorded (never a new grant), status event logged] depends_on_capabilities: [venture-leader-nomination, status-installments, resident-economic-record, track-differentiated-bar] reads_memory: [define/, define/playbooks/, define/records/residents/, define/registry/artifacts/] writes_memory: [define/records/residents/, define/records/decisions/] escalation: > Tier 2 on the venture. A declined recommendation stays in the record with the reason and a date it can be revisited. tags: [core, stage-gate, status-installment] - id: kill-and-reassignment name: Kill and reassignment description: > Run a venture kill so that it is survivable for the people on it — the thing portfolio-wide sharing was designed for (21:22). The resident authors the kill memo; economics do not move; the naming stands; the insight brief keeps their name; reassignment lands inside ten business days; what was learned is published internally inside thirty. domain: reassignment trigger: event event: validation fails or signal weakens on a venture a resident is assigned to ownership: human playbook: define/playbooks/08-venture-death-and-resident-reassignment.md inputs: [venture status, resident assignment, kill memo] outputs: [kill memo, Tulsa decision, reassignment inside 10 business days, internal learning note] depends_on_capabilities: [kill-and-reassignment, resident-record] reads_memory: [define/, define/playbooks/, define/records/residents/] writes_memory: [define/records/residents/, define/records/decisions/, define/registry/artifacts/] escalation: > Killing an incorporated venture is Tier 2. Reassignment is Tier 1, and depends on one pool across sites (stability-specs §2.8). tags: [core, engagement-closure, hardest-question] - id: decision-packet name: Bentonville to Tulsa decision packet description: > The weekly packet that turns retained authority into a known rhythm. Every Tier 2 item goes up dated; the standing answer window is TEN business days — one fortnightly review cycle; a missed window is logged rather than absorbed. C-6 CLOSED. This said five business days while playbooks 04 and 11 said ten, and the real disagreement was the FORUM, not the number: a weekly decision slot versus a fortnightly review are different operating designs. Flow's fortnightly review wins. The weekly packet survives as the SUBMISSION mechanism — weekly submission, fortnightly decision. The point is the sentence it makes possible — not "I don't have authority" but "you'll have an answer by the 14th." domain: decision-latency trigger: scheduled schedule: packet weekly; decision slot weekly; missed-window review quarterly ownership: hybrid playbook: define/playbooks/04-tulsa-decision-cadence.md inputs: [open Tier 2 items, field windows, Tier 1 commitments made this week] outputs: [weekly packet, decision dates, missed-window log, same-day Tier 1 record] depends_on_capabilities: [decision-packet-and-tier-map] reads_memory: [define/, define/intake/, define/playbooks/, define/records/decisions/, define/records/partners/] writes_memory: [define/records/decisions/] escalation: > Tier 3 — a named closing field window with what is lost attached. Rare by design; it stops working the moment it is used for convenience. tags: [core, tulsa-interface] - id: design-partner-access name: Design partner access and champion protection description: > The entry vector. Access gates discovery, ventures and everything the residency is a residency in. The competitor sentence is said in the first meeting (48:20); findings reach the champion before their leadership; a written blast-radius agreement precedes fieldwork; credit is public at spinout. Deal terms are never negotiated locally. domain: design-partner-access trigger: continuous ownership: human playbook: define/playbooks/01-design-partner-origination.md inputs: [target list, champion relationships, access status] outputs: [engaged design partner, blast-radius agreement, champion exposure notes] depends_on_capabilities: [design-partner-access, champion-protection] reads_memory: [define/, define/intake/, define/playbooks/, define/records/partners/] writes_memory: [define/records/partners/] escalation: > Terms — founding common, customer period, right to invest (49:15) — are Tier 2 and partner-owned. The room is Tier 1. tags: [core, prospect-qualification, entry-vector] - id: informant-practice name: Field practice with informants description: > The people whose workflows are the raw material sit outside every incentive structure. Commercial intent is stated in plain words, findings are returned in the moment, and nothing ships that degrades an informant's work without their knowledge. This is a quality control as much as an ethical one — the engine runs on truth from the field. domain: informant-reciprocity trigger: continuous ownership: human playbook: define/playbooks/02-ethnographic-discovery-run.md inputs: [fieldwork schedule, informant list] outputs: [consent stated, findings returned, field notes into the register] depends_on_capabilities: [informant-reciprocity] reads_memory: [define/, define/playbooks/, define/records/partners/] writes_memory: [define/registry/artifacts/] escalation: > A finding that would embarrass an informant's employer goes to the champion first under the blast-radius agreement. tags: [core, quality, extraction-risk-1] - id: artifact-register name: Artifact register description: > Every insight brief, product concept, hypothesis map, experiment plan and prototype brief — named, dated, attributed, with a resident's name where a resident authored it. In a market where a resident's title has no external ladder, a well-made named artifact is how status gets paid before a spinout exists. Quality standard belongs to Taste; the record belongs here. domain: artifact-register trigger: on_demand ownership: hybrid playbook: define/playbooks/10-artifact-production-and-review.md inputs: [artifact, author, date, publication envelope] outputs: [register entry, publication decision within the envelope] depends_on_capabilities: [artifact-register] reads_memory: [define/, define/playbooks/, define/records/partners/, define/registry/artifacts/] writes_memory: [define/registry/artifacts/] escalation: > Publication is Tier 1 inside the blast-radius and competitive-sensitivity envelope, Tier 2 at its edge. tags: [status-currency, defers-to-taste] # --- added v1.2 (Entropy) ------------------------------------------------- - id: venture-formation name: Venture formation and validation record description: > Hold the venture as a record rather than as a conversation. Playbooks 03 and 05 already produce every field — name, stage, riskiest assumption, the validation signal chosen AT INCORPORATION and not at the review (47:32–48:06), the clock, extensions, the proposed and actual formation dates, the chargeback end date, the named leader and technical lead — and nothing held them, so the Wednesday standup's core object had no home and the grant interlock had nothing to compare against. The single load-bearing field is the proposed formation date. It is the left-hand operand of the only blocking gate in the operating model (25:23, 23:08). Held as prose in a memo, that gate asserts over an empty set, is vacuously true, and renders as holding while blind. domain: venture-lifecycle trigger: event event: an insight brief clears into a concept; and again at every stage change ownership: human playbook: define/playbooks/03-concept-to-virtual-portfolio-company.md inputs: [insight brief, concept memo, validation evidence, Tulsa decisions] outputs: [venture record, proposed formation date, named validation signal, validation clock, concept_queue_depth] depends_on_capabilities: [venture-formation, venture-record] reads_memory: [define/, define/playbooks/, define/records/decisions/, define/registry/artifacts/, define/records/ventures/] writes_memory: [define/records/ventures/] escalation: > Incorporation and capital deployment are RETAINED BY TULSA (investment) and stay retained — stability-specs §9.2. This skill records the date Tulsa decided; it does not decide it, and it relocates no right. Where a formation date and an unpapered grant collide, playbook 03 DP1 governs: move the incorporation, not the grant. tags: [core, closes-G-1, interlock-operand, hard-rule] - id: field-run name: Field run record description: > One record per discovery run — dates, design partner, residents present, informants, consent stated, findings-returned date, hours to synthesis. Its purpose is arithmetic: resident floor-share is derived from these records and from nothing else, and playbook 12 calls that number the budget control, not merely a quality control. Undeclared, it was self-reported by the person whose budget it justifies and rendered to Tulsa as though read from a record. A month with no field-run record renders UNKNOWN, never a typed number. Carries informant record rights (SovereigntyMap SR-5): consent covers the record and not only the observation, credit preference is a field, and withdrawal drops the attribution while the finding survives. domain: field-cadence trigger: event event: a discovery run is scheduled, and again when it closes ownership: hybrid playbook: define/playbooks/02-ethnographic-discovery-run.md inputs: [run schedule, partner floor access, resident assignments, informant list] outputs: [field-run record, floor-days per resident, consent log, findings-returned date, informant credit_preference, withdrawn_date] depends_on_capabilities: [informant-reciprocity, design-partner-access, field-run-record] reads_memory: [define/, define/playbooks/, define/records/partners/, define/records/field-runs/] writes_memory: [define/records/field-runs/, define/registry/artifacts/] escalation: > Local. Run scoping is the OP's; method questions route to Dan Fisher as a consult, used early and visibly (decision 009, playbook 02 Level 2). The floor-share threshold is >=40% green / <30% red, set from playbook 11's run-shaped week (C-5 CLOSED). K21 is a separate, wider allocation test. tags: [core, closes-G-2, budget-control, quality] - id: spinout-formation name: Spinout and operating-team formation record description: > The destination of the whole ladder, and until now the only stage of it with no record, no writer and no surface. One record per spinout: operating-team seats and who filled them, founders-common issuance below the dotted line (37:46, 39:06, 15:01), the chargeback taper and its end date, and the design partner's credit-at-spinout as actually delivered. Credited to /fw-consonance C-8, which found the omission; declared here because it is the same act as the two above. domain: venture-lifecycle trigger: event event: a venture clears validation and an operating team is being formed ownership: human playbook: define/playbooks/09-spinout-and-operating-team-formation.md inputs: [validation decision, venture record, resident records, partner record] outputs: [spinout record, operating-team seats, chargeback taper end date, credit delivered] depends_on_capabilities: [venture-record, resident-record, champion-protection] reads_memory: [define/, define/playbooks/, define/records/ventures/, define/records/residents/, define/records/partners/] writes_memory: [define/records/spinouts/, define/records/residents/, define/records/decisions/] escalation: > Operating-team offers are RETAINED BY TULSA (hiring) — stability-specs §9.2 — and cannot be promised at intake (resident-term-sheet §4). This skill records the outcome, never commits it. tags: [core, closes-C-8, terminal-reward, credited-to-consonance] # --- end v1.2 additions --------------------------------------------------- # Track conformance (PROTOCOL.md §9.1, agency/studio): # prospect qualification -> design-partner-access # engagement delivery -> informant-practice, artifact-register # stage-gate monitoring -> readiness-review, venture-leader-naming # engagement closure -> kill-and-reassignment # Human-ownership skills present: 5 of 8. Deliberate — every judgment call in # this design (grants, naming, kills, champions, informants) stays with a person.