STD-ECON-001 — Factory Economic Gate

Purpose

Every material completed operation must answer whether the Factory is becoming cheaper and more reusable. This standard defines that gate, what counts as evidence to support it, and when it applies.

This is the completion-side complement to STD-GATE-001 (the pre-execution gate).

STD-GATE-001 asks: Is this work authorized to begin? STD-ECON-001 asks: Did this work improve the Factory's economics?

---

Central Economic Law

Every change must either improve the next run, remove unnecessary ownership, or clearly admit that it does neither.

Do not count building more Factory as Factory improvement. The proof is that future accepted outcomes require less rediscovery, less duplication, less human effort, less model spend, less risk, or less time.

---

The Gate

Every material completion receipt MUST include:

IS_THE_NEXT_RUN_GETTING_CHEAPER_AND_MORE_REUSABLE=
WHY=

This is not optional commentary. It is part of the completion receipt and is enforced by CI.

Allowed values

IS_THE_NEXT_RUN_GETTING_CHEAPER_AND_MORE_REUSABLE=YES

Only when external, observable evidence shows the change reduces future cost, rediscovery, duplication, human effort, model usage, failure risk, or execution time — or confirms reuse.

IS_THE_NEXT_RUN_GETTING_CHEAPER_AND_MORE_REUSABLE=NO

When the change is necessary but creates no compounding value. Emergency work, incident remediation, and credential rotations typically land here. NO is a correct answer.

IS_THE_NEXT_RUN_GETTING_CHEAPER_AND_MORE_REUSABLE=NOT_ESTABLISHED

When reuse or economic improvement has not yet been proven. A capability candidate has been created but not yet confirmed to compound. NOT_ESTABLISHED is a correct terminal state, not a failure or placeholder.

---

Amendment 1 — Evidence Before Claiming

Rule

IS_THE_NEXT_RUN_GETTING_CHEAPER_AND_MORE_REUSABLE=YES requires external, observable evidence. Agent assertion is not evidence. Narrating intent is not evidence.

What counts as evidence

Allowed evidence	Why it counts
A second run consumed the capability without re-deriving it	Observed reuse event
CI component invocation count across N repos (measurable delta)	Measured reduction in duplication
Formula materialized and re-run against a second estate	Cross-estate reuse confirmed
Confirmed cost/time delta with before/after measurement	Quantified improvement
Confirmed removal of prior custom implementation replaced by upstream	Ownership converged
Issue/model-call count reduction with documented baseline	Measured reduction
Upstream adoption confirmed (maintainer merge, closed issue)	Converged to upstream

What does NOT count as evidence

Invalid claim	Why it fails
"I believe this will be reused"	Assertion, not observation
"This creates reusable capability"	Output, not proof of reuse
"Code was written"	Output
"MR merged"	Delivery, not economic proof
"Formula created"	Output
"CI is green"	Correctness, not economics
"Documentation added"	Output
"Automation exists"	Output

Measurement boundary required

Every YES claim MUST state the population examined: - which repos, runs, time window, or estates were measured - what the baseline was before the change

An unmeasured improvement claim is NOT_ESTABLISHED, not YES.

---

Amendment 2 — Gate Applicability

Gate APPLIES (hard gate — CI enforces)

The gate applies to all of the following:

Engineering Standard change or creation
Formula creation or update
Order creation or update
Agent definition change
Pack change
CI component creation or update
IaC change that provisions or removes infrastructure
BluCity topology change (Rig registration, endpoint binding)
Capability promotion or removal
Security remediation Bead closure with DONE=YES
Estate inventory or audit committed to BluCity-Docs
Any MR targeting release/v0.N.x or main
Any Bead closure where DONE=YES is asserted
Handoff receipt (session end)

Gate is OPTIONAL (encouraged, not CI-enforced)

Typo or formatting fix with no logic change
Dependency version bump with no behavior change
Log or comment addition only
Test-data fixture update
Confirmed non-unique scratch file cleanup

Gate is NOT APPLICABLE

Emergency credential rotation — mark NO with explanation, not subject to CI gate
WIP commit or draft MR — gate applies at merge, not at draft creation
Read-only investigation with no state change

---

Full Completion Receipt Template

Append this to every material completion. Every field is required for material changes; unknown quantities are stated as unknown, not omitted.

REQUESTED_EFFECT=
VERIFIED=

SOURCE_DELIVERED=
RUNTIME_VERIFIED=
EVIDENCE=

CAPABILITY_CHANGE=
  NONE | CANDIDATE | UPDATED | PROVEN | REMOVED | TRANSFERRED_TO_UPSTREAM

REUSE_EFFECT=
  REUSED_EXISTING_UPSTREAM       # used an upstream capability instead of owning it
  REMOVED_CUSTOM_OWNERSHIP       # eliminated a Bluefly-owned reimplementation
  CONSOLIDATED_DUPLICATION       # merged N copies into one canonical location
  CREATED_REUSABLE_CAPABILITY    # new capability confirmed reused on ≥2 runs/estates
  IMPROVED_DISCOVERABILITY       # existing capability now found without re-derivation
  REDUCED_RE_DERIVATION          # fewer agents reconstructing established knowledge
  REDUCED_MODEL_USAGE            # fewer tokens/calls needed per outcome
  REDUCED_HUMAN_INTERVENTION     # fewer human approval loops per outcome
  REDUCED_FAILURE_RISK           # confirmed class of failures removed
  REDUCED_EXECUTION_TIME         # measured wall-clock or agent-turn reduction
  IMPROVED_VERIFICATION          # independent verification now faster or more reliable
  IMPROVED_RECOVERY              # disaster recovery path proven shorter or more reliable
  NONE                           # no economic improvement; change was necessary anyway
  NOT_ESTABLISHED                # improvement not yet measurable

REUSED_COUNT=          # number of downstream runs/repos that consumed without re-deriving
                       # (unknown if not measured — do not manufacture)
RE_DERIVED_COUNT=      # number of times this was reconstructed before this change
                       # (unknown if not measured)
NEW_WORK_COUNT=        # net new capability not previously possible
HUMAN_TIME=            # measured delta (unknown if not measured)
MODEL_COST=            # measured delta (unknown if not measured)
INFRA_COST=            # measured delta (unknown if not measured)
FAILURES=              # confirmed failure classes prevented (unknown if not measured)
REGRESSIONS=           # regressions introduced (0 must be confirmed, not assumed)

IS_THE_NEXT_RUN_GETTING_CHEAPER_AND_MORE_REUSABLE=
WHY=

WHAT_WILL_THE_NEXT_AGENT_REUSE=
WHAT_WOULD_STILL_HAVE_TO_BE_RE_DERIVED=

---

Relationship to Other Standards

Standard	Role	Relationship
STD-GATE-001	Pre-execution gate (17-point authorization vector)	Complementary — GATE-001 governs start, ECON-001 governs completion
execution-receipt-specification.md	Canonical receipt structure	ECON-001 adds the mandatory economic section
evidence-contract.md	Evidence discipline (OBSERVED / INFERRED / NOT_ESTABLISHED)	ECON-001 inherits the same evidence rules — YES requires OBSERVED evidence
capability-invariant.md	Every capability converges to smallest stable authority	ECON-001 is the measurement surface for convergence progress

---

CI Enforcement

A validate-economic-gate CI component in gitlab_components enforces this gate:

• Hard fails any MR to release/v0.N.x or main that is subject to the gate and does not contain IS_THE_NEXT_RUN_GETTING_CHEAPER_AND_MORE_REUSABLE= in the MR description or linked Bead note.
• Warns (not fails) when IS_THE_NEXT_RUN_GETTING_CHEAPER_AND_MORE_REUSABLE=YES is present but REUSE_EFFECT= is absent or NOT_ESTABLISHED.
• Skips MRs explicitly labeled gate:economic-exempt with a written justification.

Retroactive application: net-new MRs only from the introduction of this standard. Existing open MRs receive a one-release grace period (release/v0.2.x baseline).

---

Self-Applied Receipt

REQUESTED_EFFECT=Bind a mandatory economic gate to all material Factory completions,
  with clear evidence standards and applicability thresholds.
VERIFIED=NOT_YET — standard re-promoted; CI component already active (gitlab_components !1233,
  merged 2026-09-26).

SOURCE_DELIVERED=YES
RUNTIME_VERIFIED=PARTIAL — the validate-economic-gate CI component has been enforcing this
  gate estate-wide since 2026-09-26, but this standard's own canonical text did not exist at
  its documented path until this change (prior promotion MR !511 closed without merging
  2026-09-30, root cause: this file's own status: active value failed the validate:docs-
  governance STD-034 status-vocabulary check — fixed here to status: canonical).
EVIDENCE=gitlab_components!1233 (merged); blucity-docs!511 (closed, not merged); bead mba-022.

CAPABILITY_CHANGE=UPDATED

REUSE_EFFECT=REDUCED_RE_DERIVATION, IMPROVED_DISCOVERABILITY, CONSOLIDATED_DUPLICATION
  (removes the live gap between an enforced CI gate and its missing canonical definition)

REUSED_COUNT=unknown
RE_DERIVED_COUNT=unknown (estimated: every session that manually applied the gate without
  finding this file, since it was never at its canonical path)
NEW_WORK_COUNT=0 (re-promoting existing drafted content, not new policy)
HUMAN_TIME=unknown
MODEL_COST=unknown
INFRA_COST=0

IS_THE_NEXT_RUN_GETTING_CHEAPER_AND_MORE_REUSABLE=NOT_ESTABLISHED
WHY=The standard is now at its canonical path with a status value the governance validator
  accepts. No downstream agent has yet consumed it from this location. Updates to YES once a
  confirmed downstream use is observed without re-derivation.

WHAT_WILL_THE_NEXT_AGENT_REUSE=The STD-ECON-001 receipt template, the evidence
  threshold (Amendment 1), and the gate applicability list (Amendment 2), from its
  canonical path instead of ledger/.

WHAT_WOULD_STILL_HAVE_TO_BE_RE_DERIVED=Nothing for this standard. BLU-BIBLE re-indexing
  (bead mba-022, after this merges) remains separate.