Engagement repo bootstrap (private; proof-plan skeleton; fleet-onboard)

[ojfbot]

Question

(Task — manual unblocking work.) Create the private engagement repo named per the northstar ruling, seeded with the proof-plan skeleton (decision log, baseline ledger, eval ledger, plumbing journal, sanitized case-study skeleton with blanks), and register it on every fleet surface via /fleet-onboard.

Type

task

Context

Map: #344. Skeleton spec: vault wiki/synthesis/texas-rr-engagement-brief §Proof plan. Client material stays private; the repo is the dossier-in-progress.

Blocked by

#345 (Engagement northstar — naming ruling names the repo).

Resolution (filled at close)

—

[ojfbot]

Unblocked by #345 (closed 2026-08-01). Bootstrap inputs from the ruling:

• Repo name: fieldwork-1 (private). Slug l1-fieldwork-1, registered in decisions/northstar/README.md (PR wayfinder #345: engagement northstar — l1-fieldwork-1 registered #357) with path: ../fieldwork-1/.claude/northstar.md — landing the file below closes the absent-checkout lint WARN.
• At bootstrap, additionally: (1) insert client-named baseline numbers into the northstar file INSIDE the private repo (they stay out of core — client-data ruling, see also Public-map client-intel sanitization #356); (2) execute the map re-anchor ruled at Engagement northstar: naming, properties, movement contract #345: flip decisions/wayfinder/texas-rr-engagement.md frontmatter northstar: to l1-fieldwork-1 once ns:l1-fieldwork-1#P1 resolves (resolve-or-fail — verify with resolve-anchor.mjs before writing).
• Draft below is client-light on purpose (this issue is public). Verbatim start point for fieldwork-1/.claude/northstar.md:

---
type: northstar
slug: l1-fieldwork-1
tier: L1
app: fieldwork-1
ladders_up_to: l2-ojfbot
status: active
# Ruled at wayfinder #345 close (2026-08-01); map: core/decisions/wayfinder/texas-rr-engagement.md.
# Engagement-scoped + ordinal by design: this northstar retires at sustainment/exit; a second
# engagement mints l1-fieldwork-2. NUMERIC CURRENTS/TARGETS ARE PROPOSALS (operator to calibrate).
# Client-named baseline numbers are inserted HERE (private repo) at bootstrap — never in core.
properties:
  - id: P1
    name: "The full engagement arc is traversed"
    target: "lead → research → discovery → scoped delivery → sustainment, each stage passed through its entrance/success gate (gates ruled at core#350)."
    current: 15
    verification: "Stage-gate ledger in this repo: per-stage entrance/success evidence, dated. Current 15% = lead + research substantially done (vault brief, entity ledger, ground-truth pass; competitor teardown core#348 still open); discovery not begun."
    ladders_up_to: "ns:l2-ojfbot#P1"
    okr_drivers: []
  - id: P2
    name: "Value lands in the customer's numbers"
    target: "Every shipped slice carries a baseline→after delta stated in the customer's units (hours/week, cycle time, % complete files, retained relationships — never tokens) and accepted by the customer. First movement = ≥1 workflow absorbed with a measured delta."
    current: 0
    verification: "Baseline ledger (proof-plan artifact 2) + measured-delta entries, each traceable to a shipped slice."
    ladders_up_to: "ns:l2-ojfbot#P1"
    okr_drivers: []
  - id: P3
    name: "The dossier is genre-complete (G1–G6)"
    target: "All six FDE deliverable genres (core/decisions/research/2026-08-01-fde-deliverables-audit.md) evidenced by real artifacts in this repo; anything leaving the repo passes the sanitization gate (core#347) first."
    current: 0
    verification: "Per-genre artifact index in the dossier; each genre row points at ≥1 real artifact."
    ladders_up_to: "ns:l2-ojfbot#P1"
    okr_drivers: []
  - id: P4
    name: "The playbook shadow fills per completed stage"
    target: "Each completed stage has its pattern captured after the evidence exists — never ahead of it; at exit the shadow covers the full arc and the scale hypothesis's falsifiable predicates (≥60% reuse, ≤2 weeks to first demo) are testable."
    current: 0
    verification: "Playbook-shadow doc (format = core#351's ruling) with per-stage entries dated after stage completion."
    ladders_up_to: "ns:l2-ojfbot#P2"
    okr_drivers: []
  - id: P5
    name: "Delivery is mechanized and auditable (recursive SE)"
    target: "Every recurring engagement workflow/component built for delivery runs as an automated, audited pipeline — requirements→design→verification traceable per the NASA SEH process groups (system design top-down, product realization bottom-up, technical management cross-cutting) — with manual effort measured and trending down; auditability/transparency/accountability/accuracy are the quality bar. (17-process mapping detail = core#350, widened scope.)"
    current: 0
    verification: "Component inventory with per-component trace (requirement → design record → verification evidence) + a manual-vs-automated effort ledger."
    ladders_up_to: "ns:l2-ojfbot#P2"
    okr_drivers: []
---

# Northstar — fieldwork-1 (L1)

**Vision.** The first self-funded field engagement traverses the full FDE arc — lead → research →
discovery → scoped delivery → sustainment — on the fleet's own SE engine, with value measured in
the customer's numbers, a genre-complete proof-of-work dossier filling as the work happens (never
reconstructed), and a playbook shadow capturing patterns only after the evidence exists.
Over-delivery is the strategy; **unmeasured** over-delivery is the failure mode. The standing
boundary: absorb workflows, never the role — each absorbed workflow gets a name, a baseline, and a
metric, or it goes to the backlog. The engagement's own machinery is itself SE-governed (P5):
components built to serve delivery are defined top-down, realized bottom-up, and cross-cut by
technical management, biased to mechanical automation over manual effort. Compensation is the
documented engagement itself — the dossier is the deliverable the fleet keeps. Boundary rules and
pain hypotheses live in the vault brief (selfco `wiki/synthesis/texas-rr-engagement-brief`), cited
here rather than restated.

**Movement contract.** Currents are hand-asserted by the operator only. Slice sessions NEVER write
`current:` — every engagement PR body carries
`Movement proposal: ns:l1-fieldwork-1#Pn N% -> M% — evidence: …`; `record-movement.mjs` appends
`status.jsonl` from merged PRs only. A slice that cannot name its property does not dispatch
(work with no metric attached doesn't start).

## P1 — The full engagement arc is traversed

The single arc-% signal for the standup. Fed by the stage-gate ledger (#350's gates) — the
property is the rollup the ledger feeds, same as every other northstar. Advances `l2-ojfbot#P1`
by shipping a usable surface in its natural venue: the customer's environment.

## P2 — Value lands in the customer's numbers

The anti-odd-jobs axis. No token economics, no vanity metrics — deltas in the units the customer
already uses, accepted by the customer. Advances `l2-ojfbot#P1`: value shown where the work lives.

## P3 — The dossier is genre-complete (G1–G6)

The compensation. G3 (full deployment arc) is the fleet's confirmed #1 gap; this engagement is
the instrument that closes it. Sanitization-gated before anything reaches the presence map.

## P4 — The playbook shadow fills per completed stage

The scale thesis (G5), kept honest: pattern capture strictly trails evidence. Advances
`l2-ojfbot#P2` — it is the mechanism that makes engagement work repeatable and legible.

## P5 — Delivery is mechanized and auditable (recursive SE)

The operator-elevated axis (2026-08-01): the NASA SEH process model applied to the components
built for delivery — least manual effort, most mechanical automation, auditability, transparency,
accountability, accuracy. Distinct from the client-facing stage naming (literal ISO-15288 framing
stays ruled out). Advances `l2-ojfbot#P2`: the engagement's own work becomes self-measuring.