<?xml version="1.0" encoding="utf-8"?>
<entry xmlns="http://www.w3.org/2005/Atom" xml:base="https://docs.spec-kitty.ai/adr/3.x/2026-07-14-2-doctrine-to-core-mission-type-resolution-unification/">
  <id>https://docs.spec-kitty.ai/adr/3.x/2026-07-14-2-doctrine-to-core-mission-type-resolution-unification/</id>
  <title>ADR: Doctrine → Charter → Core Mission-Type Resolution Unification (governance first)</title>
  <updated>2026-07-14T00:00:00.000Z</updated>
  <author><name>Spec Kitty</name></author>
  <link rel="alternate" type="text/html" href="https://docs.spec-kitty.ai/adr/3.x/2026-07-14-2-doctrine-to-core-mission-type-resolution-unification/"/>
  <summary>`resolve_mission_type_context` becomes the one charter-mediated mission-type resolver; an unknown type hard-errors instead of defaulting or being inferred from `template_set`.</summary>
  <content type="html">&lt;h2 id=&quot;context-and-problem-statement&quot;&gt;Context and Problem Statement&lt;/h2&gt;
&lt;p&gt;Spec Kitty resolves per-mission-type behaviour — governance, templates, gates,
step contracts — through &lt;strong&gt;two parallel trees and three competing governance
surfaces&lt;/strong&gt;, with a hardcoded &lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt; bias woven through the core loop.&lt;/p&gt;
&lt;p&gt;The two trees:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;code dir=&quot;auto&quot;&gt;src/doctrine/missions/&amp;#x3C;type&gt;/&lt;/code&gt; — the canonical catalogue. It is already the
source of truth: &lt;code dir=&quot;auto&quot;&gt;MissionTemplateRepository.default_missions_root()&lt;/code&gt;
(&lt;code dir=&quot;auto&quot;&gt;doctrine/missions/repository.py:98-110&lt;/code&gt;) resolves here, and
&lt;code dir=&quot;auto&quot;&gt;template/manager.py:88&lt;/code&gt; copies it into the deployed &lt;code dir=&quot;auto&quot;&gt;.kittify/missions/&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;&lt;code dir=&quot;auto&quot;&gt;src/specify_cli/missions/&amp;#x3C;type&gt;/&lt;/code&gt; — &lt;strong&gt;derived copies&lt;/strong&gt; that several core
readers still bind to directly (dossier &lt;code dir=&quot;auto&quot;&gt;ManifestRegistry.load_manifest&lt;/code&gt;,
&lt;code dir=&quot;auto&quot;&gt;dossier/manifest.py:163-193&lt;/code&gt;; &lt;code dir=&quot;auto&quot;&gt;_packaged_missions_dir&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;mission.py&lt;/code&gt;; template
resolution, &lt;code dir=&quot;auto&quot;&gt;core/project_resolver.py&lt;/code&gt;). The two trees have already drifted with
no parity guard (see #2628) — and the drift runs &lt;strong&gt;specify_cli-ahead&lt;/strong&gt;: the
&lt;code dir=&quot;auto&quot;&gt;specify_cli&lt;/code&gt; copies carry live entries (e.g. &lt;code dir=&quot;auto&quot;&gt;runtime.charter-lint.decay&lt;/code&gt; →
&lt;code dir=&quot;auto&quot;&gt;lint-report.json&lt;/code&gt;, a &lt;code dir=&quot;auto&quot;&gt;blocking: false&lt;/code&gt; flag) that the doctrine copies lack.
The doctrine-side &lt;code dir=&quot;auto&quot;&gt;expected-artifacts.yaml&lt;/code&gt; reader (&lt;code dir=&quot;auto&quot;&gt;repository.py:304&lt;/code&gt;) exists
but is &lt;strong&gt;dead&lt;/strong&gt; (tests-only).&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;The three governance surfaces per mission type:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;code dir=&quot;auto&quot;&gt;mission_types/&amp;#x3C;type&gt;.yaml&lt;/code&gt; &lt;code dir=&quot;auto&quot;&gt;governance_refs&lt;/code&gt; — &lt;strong&gt;inert&lt;/strong&gt; (no runtime reader;
only CLI display) and &lt;strong&gt;dangling&lt;/strong&gt;: &lt;code dir=&quot;auto&quot;&gt;software-dev.yaml:11-12&lt;/code&gt; references
&lt;code dir=&quot;auto&quot;&gt;DIR-010/DIR-011&lt;/code&gt;, which resolve to nothing (real ids are &lt;code dir=&quot;auto&quot;&gt;DIRECTIVE_0NN&lt;/code&gt;).&lt;/li&gt;
&lt;li&gt;&lt;code dir=&quot;auto&quot;&gt;missions/&amp;#x3C;type&gt;/governance-profile.yaml&lt;/code&gt; &lt;code dir=&quot;auto&quot;&gt;selected_*&lt;/code&gt; — live via
&lt;code dir=&quot;auto&quot;&gt;prompt_builder.py:346&lt;/code&gt;, correctly keyed off &lt;code dir=&quot;auto&quot;&gt;meta.json mission_type&lt;/code&gt;,
&lt;strong&gt;hard-fails&lt;/strong&gt; rather than falling back — but ships &lt;strong&gt;empty&lt;/strong&gt; for all four types.&lt;/li&gt;
&lt;li&gt;&lt;code dir=&quot;auto&quot;&gt;missions/&amp;#x3C;type&gt;/actions/&amp;#x3C;action&gt;/index.yaml&lt;/code&gt; — live, populated, action-grained,
generating DRG &lt;code dir=&quot;auto&quot;&gt;scope&lt;/code&gt; edges consumed by &lt;code dir=&quot;auto&quot;&gt;charter context --action&lt;/code&gt;.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;The concrete defect this creates: the action-scoped path infers mission type as
&lt;code dir=&quot;auto&quot;&gt;(template_set or &quot;software-dev-default&quot;).removesuffix(&quot;-default&quot;)&lt;/code&gt; in
&lt;code dir=&quot;auto&quot;&gt;_load_action_doctrine_bundle&lt;/code&gt; (&lt;code dir=&quot;auto&quot;&gt;charter/context.py:865&lt;/code&gt;, reached from both
&lt;code dir=&quot;auto&quot;&gt;build_charter_context:252&lt;/code&gt; and &lt;code dir=&quot;auto&quot;&gt;build_charter_context_json:3254&lt;/code&gt;). A
documentation, research, or plan mission that never set &lt;code dir=&quot;auto&quot;&gt;template_set&lt;/code&gt; silently
loads &lt;strong&gt;software-dev&lt;/strong&gt; doctrine. This is exactly the failure #883 exists to close:
non-software missions inheriting test-first/implementation/code-review doctrine by
default. A &lt;strong&gt;second incarnation&lt;/strong&gt; of the same default lives in
&lt;code dir=&quot;auto&quot;&gt;specify_cli/mission.py:575&lt;/code&gt; (&lt;code dir=&quot;auto&quot;&gt;get_mission_type&lt;/code&gt; returns &lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt; when
&lt;code dir=&quot;auto&quot;&gt;meta.json&lt;/code&gt; omits the key), and it feeds the &lt;strong&gt;dossier&lt;/strong&gt; path — so the leak has
two surfaces, and one of them (dossier) sits behind a package boundary the charter
layer may not import.&lt;/p&gt;
&lt;p&gt;Underlying all of this is a north-star intent the team has now made explicit: the
&lt;code dir=&quot;auto&quot;&gt;src/specify_cli/missions/&lt;/code&gt; tree should eventually be &lt;strong&gt;deleted&lt;/strong&gt;, and
&lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt; should become an ordinary built-in &lt;strong&gt;doctrine&lt;/strong&gt; mission type on
equal footing with &lt;code dir=&quot;auto&quot;&gt;documentation&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;research&lt;/code&gt;, and &lt;code dir=&quot;auto&quot;&gt;plan&lt;/code&gt; — feeding step
contracts, gates, and templates from the doctrine, &lt;strong&gt;through the charter&lt;/strong&gt;, into
the execution loop, with no &lt;code dir=&quot;auto&quot;&gt;software-dev-default&lt;/code&gt; special-casing and no
hardcoded core knowledge of &lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt;.&lt;/p&gt;
&lt;p&gt;This ADR records the architecture that gets us there and the decision to land
mission-type &lt;strong&gt;governance&lt;/strong&gt; as its first proven slice (issue #883). Its claims and
work breakdown were pressure-tested by an adversarial four-lens second-opinion
review before acceptance; the review’s corrections are folded in below.&lt;/p&gt;
&lt;h2 id=&quot;decision-drivers&quot;&gt;Decision Drivers&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Single canonical authority (DIRECTIVE_044).&lt;/strong&gt; Three governance surfaces and
two mission trees violate single-source; a fix that adds a fourth surface or a
“keep the trees in sync” guard entrenches the split instead of removing it.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Close the software-dev-default leak by construction (DIRECTIVE_043)&lt;/strong&gt; — on
every surface that carries it, not only the prompt path.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Layered, mission-aware governance (issue #883, epic #461, #832).&lt;/strong&gt; Governance
must resolve as &lt;code dir=&quot;auto&quot;&gt;project_charter ⊕ shipped_mission_type ⊕ project_override&lt;/code&gt;,
keyed off &lt;code dir=&quot;auto&quot;&gt;meta.json mission_type&lt;/code&gt;, overridable in #832’s shipped → org →
project order.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Extend, do not invent (#2628 lesson).&lt;/strong&gt; Reuse existing doctrine machinery
rather than parallel declaration/merge mechanisms — while being honest about
where “reuse” still requires real adapter work.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;A widening seam, not a governance dead-end.&lt;/strong&gt; Doctrine &lt;em&gt;offers&lt;/em&gt;, charter
&lt;em&gt;activates and customises&lt;/em&gt;, core (an FSM “semi-prepared for states &amp;#x26; transitions
as config”) &lt;em&gt;consumes&lt;/em&gt;. The resolution seam #883 builds must be the general path
templates, gates, and step contracts flow through next.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The doctrine path goes live; the &lt;code dir=&quot;auto&quot;&gt;specify_cli&lt;/code&gt; path is removed — a transparent
swap.&lt;/strong&gt; The dead doctrine read-paths become the live source of truth and the
derived &lt;code dir=&quot;auto&quot;&gt;specify_cli/missions/&lt;/code&gt; copies are deleted, with behaviour preserved so
the user never notices the swap. Preservation is verified by &lt;strong&gt;transitional&lt;/strong&gt;
parity scaffolding that is added at the start of the migration and &lt;strong&gt;deleted
before merge&lt;/strong&gt; — not by a surviving parity gate. Enduring verification is
behavioural, at the doctrine-module and integration level.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;No code kept solely to avoid test churn.&lt;/strong&gt; Tests are expected to change
substantially; shims, compat wrappers, and signatures preserved only to keep old
callers green are anti-patterns here. The suite is updated to assert the new
behaviour rather than the code contorted to protect the old suite.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The &lt;code dir=&quot;auto&quot;&gt;MissionType&lt;/code&gt; doctrine artefact becomes load-bearing.&lt;/strong&gt; It is enhanced to
be the single doctrine answer to “what is this mission type, what steps does it
contain, what gates are checked?” — its steps (action sequence / step contracts),
its gates (expected-artifacts), and its governance all resolve &lt;em&gt;through&lt;/em&gt; it,
doctrine → charter → core. It stops being a mostly-inert descriptor.&lt;/li&gt;
&lt;/ul&gt;
&lt;h2 id=&quot;considered-options&quot;&gt;Considered Options&lt;/h2&gt;
&lt;ol&gt;
&lt;li&gt;&lt;strong&gt;Populate &lt;code dir=&quot;auto&quot;&gt;governance_refs&lt;/code&gt;&lt;/strong&gt; (the 2026-07-04 triage’s suggestion).&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Fill &lt;code dir=&quot;auto&quot;&gt;governance-profile.yaml&lt;/code&gt; only&lt;/strong&gt;, leaving the action-scoped leak in place.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;A unified &lt;code dir=&quot;auto&quot;&gt;doctrine → charter → core&lt;/code&gt; resolution seam&lt;/strong&gt; — one charter-mediated
resolver keyed off &lt;code dir=&quot;auto&quot;&gt;meta.json&lt;/code&gt;, feeding both consumers; governance as the first
slot of a bundle whose later slots are templates/gates/step contracts;
&lt;code dir=&quot;auto&quot;&gt;governance_refs&lt;/code&gt; retired; the leak closed; &lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt; demoted to a peer —
&lt;strong&gt;chosen&lt;/strong&gt;, delivered as slice 1 of a &lt;code dir=&quot;auto&quot;&gt;specify_cli/missions&lt;/code&gt; retirement epic.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Full dual-tree merge now.&lt;/strong&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;h2 id=&quot;decision-outcome&quot;&gt;Decision Outcome&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Chosen option:&lt;/strong&gt; “A unified &lt;code dir=&quot;auto&quot;&gt;doctrine → charter → core&lt;/code&gt; resolution seam”
(Option 3), delivered as &lt;strong&gt;slice 1 of a &lt;code dir=&quot;auto&quot;&gt;specify_cli/missions&lt;/code&gt; retirement epic&lt;/strong&gt;.&lt;/p&gt;
&lt;h3 id=&quot;decisions-recorded&quot;&gt;Decisions recorded&lt;/h3&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;One charter-mediated resolver.&lt;/strong&gt;
&lt;code dir=&quot;auto&quot;&gt;charter.mission_type_profiles.resolve_mission_type_context(repo_root, *, mission_type=None, feature_dir=None) -&gt; ResolvedMissionType&lt;/code&gt; is the single entry
point for per-mission-type resolution. It resolves the type from an explicit
argument → &lt;code dir=&quot;auto&quot;&gt;feature_dir/meta.json&lt;/code&gt; → &lt;strong&gt;hard error for an &lt;em&gt;unknown&lt;/em&gt; type&lt;/strong&gt;. It
never infers the type from &lt;code dir=&quot;auto&quot;&gt;template_set&lt;/code&gt; and never defaults to &lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt;.
It subsumes three functions already keyed by mission type
(&lt;code dir=&quot;auto&quot;&gt;resolve_action_sequence:266&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;resolve_mission_type_governance:325&lt;/code&gt;,
&lt;code dir=&quot;auto&quot;&gt;load_profile:175&lt;/code&gt;); their live callers (including the 5 FSM-path callers of
&lt;code dir=&quot;auto&quot;&gt;resolve_action_sequence&lt;/code&gt;) are &lt;strong&gt;migrated onto the resolver&lt;/strong&gt;. A former function
is kept only where it is genuinely the right long-term API, never as a wrapper
retained solely to avoid editing callers. Their &lt;strong&gt;two distinct hard-fail
policies&lt;/strong&gt; — action-sequence validates against the activation set with no escape
hatch; governance tolerates an unknown type when a project override exists
(&lt;code dir=&quot;auto&quot;&gt;:392&lt;/code&gt;) — are preserved as explicit branches, not flattened.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Known-type-with-empty-grain is legitimate, not an error.&lt;/strong&gt; &lt;code dir=&quot;auto&quot;&gt;plan&lt;/code&gt; ships no
action indices and an empty type-grain today; a known mission type resolving to
an empty governance set is valid. Only an &lt;em&gt;unknown&lt;/em&gt; mission type hard-fails.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;A bundle sourced from the load-bearing &lt;code dir=&quot;auto&quot;&gt;MissionType&lt;/code&gt; artefact.&lt;/strong&gt;
&lt;code dir=&quot;auto&quot;&gt;ResolvedMissionType&lt;/code&gt; carries populated &lt;code dir=&quot;auto&quot;&gt;governance&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;action_sequence&lt;/code&gt;,
&lt;code dir=&quot;auto&quot;&gt;expected_artifacts&lt;/code&gt; (gates), and &lt;code dir=&quot;auto&quot;&gt;step_contracts&lt;/code&gt; (steps) slots in slice 1; the
&lt;code dir=&quot;auto&quot;&gt;template_set&lt;/code&gt; slot is populated in a later slice. All slots are sourced from the
doctrine &lt;code dir=&quot;auto&quot;&gt;MissionType&lt;/code&gt; artefact and the files it references (a sibling
&lt;code dir=&quot;auto&quot;&gt;governance-profile.yaml&lt;/code&gt;, the action indices, the step contracts, and
expected-artifacts), never from the &lt;code dir=&quot;auto&quot;&gt;specify_cli&lt;/code&gt; copies. The bundle is how the
artefact becomes load-bearing at the core. &lt;strong&gt;(Operator decision Q3: slice 1
covers governance + gates + steps; templates follow.)&lt;/strong&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Ordered, structured output.&lt;/strong&gt; Governance resolves to a structured
&lt;code dir=&quot;auto&quot;&gt;ResolvedGovernance&lt;/code&gt; whose selections are &lt;strong&gt;ordered&lt;/strong&gt; (list-backed with an
explicit, tested sort — not sets), so rendering is deterministic. Determinism is
a correctness property of the resolver (verified by a doctrine-module test), not a
reason to freeze a byte-snapshot of the old path forever.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Both live consumers converge on the resolver.&lt;/strong&gt; The rewire targets the &lt;em&gt;live&lt;/em&gt;
code: &lt;code dir=&quot;auto&quot;&gt;prompt_builder.py:346&lt;/code&gt; (Surface B) and &lt;code dir=&quot;auto&quot;&gt;_load_action_doctrine_bundle&lt;/code&gt;
(&lt;code dir=&quot;auto&quot;&gt;context.py:865&lt;/code&gt;, reached via &lt;code dir=&quot;auto&quot;&gt;build_charter_context:252&lt;/code&gt; &lt;strong&gt;and&lt;/strong&gt;
&lt;code dir=&quot;auto&quot;&gt;build_charter_context_json:3254&lt;/code&gt;). The dead &lt;code dir=&quot;auto&quot;&gt;_render_action_scoped:1500&lt;/code&gt; /
&lt;code dir=&quot;auto&quot;&gt;_append_action_doctrine_lines:1451&lt;/code&gt; pair and its lone test are &lt;strong&gt;deleted&lt;/strong&gt;, not
threaded.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Close the leak by threading the real mission type — with per-entry behaviour.&lt;/strong&gt;
&lt;code dir=&quot;auto&quot;&gt;mission_type&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;feature_dir&lt;/code&gt; is threaded into the two live &lt;code dir=&quot;auto&quot;&gt;build_charter_context*&lt;/code&gt;
entry points; the ~40 test callers are &lt;strong&gt;updated to the new signature&lt;/strong&gt; (not
shielded by a compat overload kept solely to avoid editing them). Behaviour is
defined &lt;strong&gt;per entry&lt;/strong&gt;, not as a blanket hard-fail: the prompt path supplies
&lt;code dir=&quot;auto&quot;&gt;feature_dir&lt;/code&gt;; planning-from-root (&lt;code dir=&quot;auto&quot;&gt;charter/context.py:90&lt;/code&gt;) requires an explicit
&lt;code dir=&quot;auto&quot;&gt;--mission-type&lt;/code&gt;; the genuinely mission-less callers (dispatch &lt;code dir=&quot;auto&quot;&gt;executor.py:270&lt;/code&gt;,
and &lt;code dir=&quot;auto&quot;&gt;workflow.py:675&lt;/code&gt;, which already degrades to &lt;code dir=&quot;auto&quot;&gt;&quot;Governance: unavailable&quot;&lt;/code&gt;) get
a defined neutral/degrade path, never a silent software-dev load.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;A single mission-type canonicalizer (&lt;code dir=&quot;auto&quot;&gt;WP-CANON&lt;/code&gt;), across the package boundary.&lt;/strong&gt;
The type key is read raw in at least two places — &lt;code dir=&quot;auto&quot;&gt;mission_type_profiles.py:380&lt;/code&gt;
and &lt;code dir=&quot;auto&quot;&gt;specify_cli/mission.py:575&lt;/code&gt; (the latter silently defaulting to
&lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt;). Because &lt;code dir=&quot;auto&quot;&gt;charter/&lt;/code&gt; may not import &lt;code dir=&quot;auto&quot;&gt;specify_cli&lt;/code&gt; (layer rule,
&lt;code dir=&quot;auto&quot;&gt;test_layer_rules.py&lt;/code&gt;), the canonicalizer lives where both may consume it, and
&lt;code dir=&quot;auto&quot;&gt;get_mission_type&lt;/code&gt;’s software-dev default is removed. Until this lands, the leak
is closed on the prompt/action-doctrine path but &lt;strong&gt;remains live on the dossier
path&lt;/strong&gt;; the ADR does not claim otherwise.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The inert &lt;code dir=&quot;auto&quot;&gt;governance_refs&lt;/code&gt; field is replaced, not merely dropped.&lt;/strong&gt; The
hand-authored, dangling &lt;code dir=&quot;auto&quot;&gt;governance_refs&lt;/code&gt; (&lt;code dir=&quot;auto&quot;&gt;models.py:186&lt;/code&gt;, CLI display
&lt;code dir=&quot;auto&quot;&gt;mission_type.py:1486,1505&lt;/code&gt;, dangling &lt;code dir=&quot;auto&quot;&gt;software-dev.yaml:11-12&lt;/code&gt;) is removed —
but the &lt;em&gt;intent&lt;/em&gt; it gestured at (a mission type declaring its governance) is
fulfilled properly by the load-bearing artefact: governance resolves through the
charter-mediated resolver keyed by the type, DRG-resolvable, no danglers.
&lt;strong&gt;(Operator decision Q1: the &lt;code dir=&quot;auto&quot;&gt;MissionType&lt;/code&gt; artefact &lt;em&gt;references&lt;/em&gt; the live sibling
&lt;code dir=&quot;auto&quot;&gt;governance-profile.yaml&lt;/code&gt; for type-grain governance — the schema’d, hard-failing
surface the resolver already reads — rather than absorbing governance as a field;
governance-profile.yaml is not retired.)&lt;/strong&gt; This removes the inert field and its
authored content updates the tests that asserted it
(&lt;code dir=&quot;auto&quot;&gt;test_mission_type_repository.py:44&lt;/code&gt;,
&lt;code dir=&quot;auto&quot;&gt;test_charter_mission_type_commands.py:242,247&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;test_activation_filtered_drg.py&lt;/code&gt;
docstring, the &lt;code dir=&quot;auto&quot;&gt;drg.py:169&lt;/code&gt; comment) to the new behaviour — the tests move with
the behaviour, they are not preserved to keep the field alive.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Two governance grains, unioned, URN-normalized disjointness.&lt;/strong&gt;
&lt;code dir=&quot;auto&quot;&gt;actions/&amp;#x3C;action&gt;/index.yaml&lt;/code&gt; is the canonical &lt;strong&gt;action-grain&lt;/strong&gt; (already live +
DRG scope edges, config-stem ids); &lt;code dir=&quot;auto&quot;&gt;governance-profile.yaml selected_*&lt;/code&gt; is the
canonical &lt;strong&gt;type-grain&lt;/strong&gt; (and the project-override target). The resolver unions
and de-dupes them; a guard forbids the same artifact in both grains, &lt;strong&gt;after
normalizing both grains to canonical URNs&lt;/strong&gt; (string equality across &lt;code dir=&quot;auto&quot;&gt;003-…&lt;/code&gt; vs
&lt;code dir=&quot;auto&quot;&gt;DIRECTIVE_003&lt;/code&gt; vs a URN gives false assurance). This is “two mechanisms with a
canonical-URN disjointness guard,” not a claim of a single authored surface.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Per-type project override — via the overlay stack, with the real adapter cost
named.&lt;/strong&gt; The override lives at
&lt;code dir=&quot;auto&quot;&gt;.kittify/doctrine/mission_types/&amp;#x3C;type&gt;/governance-profile.yaml&lt;/code&gt; and rides the
existing &lt;code dir=&quot;auto&quot;&gt;doctrine/base.py&lt;/code&gt; builtin → org → project overlay loader
(&lt;code dir=&quot;auto&quot;&gt;base.py:213-311&lt;/code&gt;) with its field-merge and &lt;code dir=&quot;auto&quot;&gt;DoctrineLayerCollisionWarning&lt;/code&gt;.
This is &lt;strong&gt;not free&lt;/strong&gt;: &lt;code dir=&quot;auto&quot;&gt;base.py:249&lt;/code&gt; keys overlays on an &lt;code dir=&quot;auto&quot;&gt;id&lt;/code&gt; field and skips
files without one, and &lt;code dir=&quot;auto&quot;&gt;governance-profile.yaml&lt;/code&gt; keys on &lt;code dir=&quot;auto&quot;&gt;mission_type&lt;/code&gt;.
&lt;strong&gt;(Operator decision Q2: ride the overlay stack.)&lt;/strong&gt; &lt;code dir=&quot;auto&quot;&gt;WP-OVERRIDE&lt;/code&gt; adds an &lt;code dir=&quot;auto&quot;&gt;id&lt;/code&gt; to
&lt;code dir=&quot;auto&quot;&gt;MissionTypeProfile&lt;/code&gt; (a coordinated &lt;code dir=&quot;auto&quot;&gt;extra=&quot;forbid&quot;&lt;/code&gt; change across the built-in
files) plus a &lt;code dir=&quot;auto&quot;&gt;BaseDoctrineRepository[MissionTypeProfile]&lt;/code&gt; subclass, so the
override flows through the same builtin → org → project loader every other
doctrine artefact uses — giving #832 org-layer support and collision auditing for
free once the adapter exists. A bespoke second field-merge in the resolver was
rejected to avoid a duplicate merge site. &lt;code dir=&quot;auto&quot;&gt;WP-OVERRIDE&lt;/code&gt; owns the adapter and ships
its own precedence + collision test.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Governance is a sibling of the state list, not a property of it.&lt;/strong&gt; Keyed off
the same &lt;code dir=&quot;auto&quot;&gt;mission_type&lt;/code&gt; the FSM uses for &lt;code dir=&quot;auto&quot;&gt;action_sequence&lt;/code&gt; (&lt;code dir=&quot;auto&quot;&gt;decision.py:601&lt;/code&gt;,
&lt;code dir=&quot;auto&quot;&gt;runtime_bridge_composition.py:186&lt;/code&gt;), never attached to individual state
definitions — so the two config axes stay separable when full “states as config”
lands.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;&lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt; becomes a peer doctrine type&lt;/strong&gt; with no new authoring; its
effective governance is &lt;strong&gt;preserved so the user does not notice the swap&lt;/strong&gt; —
verified by &lt;em&gt;transitional&lt;/em&gt; parity scaffolding (deleted before merge), not locked
by a surviving gate. Only documentation, research, and plan gain content. The
surviving &lt;code dir=&quot;auto&quot;&gt;meta.json&lt;/code&gt;-less fallback to &lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt; in &lt;code dir=&quot;auto&quot;&gt;mission.py&lt;/code&gt; is scoped to
&lt;strong&gt;template-file selection&lt;/strong&gt;, not governance, after &lt;code dir=&quot;auto&quot;&gt;WP-CANON&lt;/code&gt;.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3 id=&quot;corollaries-scope-boundaries&quot;&gt;Corollaries (scope boundaries)&lt;/h3&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;#883 delivers slice 1, not the whole retirement.&lt;/strong&gt; In scope: &lt;code dir=&quot;auto&quot;&gt;WP-CANON&lt;/code&gt;, the
resolver seam, the leak fix, &lt;code dir=&quot;auto&quot;&gt;governance_refs&lt;/code&gt; retirement, the per-type override
(ridden through the overlay stack), authored governance for
documentation/research/plan, &lt;strong&gt;step-contract resolution through the artefact
(Q3)&lt;/strong&gt;, one &lt;strong&gt;gated&lt;/strong&gt; reader migration for gates (see below), and the enforcement
gates. Out of scope, left as clean seams: &lt;strong&gt;template resolution&lt;/strong&gt; (&lt;code dir=&quot;auto&quot;&gt;template_set&lt;/code&gt;
slot), enumeration / &lt;code dir=&quot;auto&quot;&gt;mission-runtime&lt;/code&gt;, removing the &lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt; fallback in
&lt;code dir=&quot;auto&quot;&gt;mission.py:466,469&lt;/code&gt;, and finally deleting the copy step
(&lt;code dir=&quot;auto&quot;&gt;template/manager.py:88&lt;/code&gt;) and the &lt;code dir=&quot;auto&quot;&gt;specify_cli/missions/&lt;/code&gt; tree.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The dossier migration is a gated, isolated swap (operator decision).&lt;/strong&gt; The dead
doctrine reader goes live and the &lt;code dir=&quot;auto&quot;&gt;specify_cli&lt;/code&gt; copies are removed, user-invisibly.
It runs as a detachable lane, &lt;strong&gt;non-blocking for the enforcement gate&lt;/strong&gt;: (1)
reconcile the drifted &lt;code dir=&quot;auto&quot;&gt;expected-artifacts.yaml&lt;/code&gt; &lt;strong&gt;upward&lt;/strong&gt; into the doctrine tree
(porting the &lt;code dir=&quot;auto&quot;&gt;runtime.charter-lint.decay&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;lint-report&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;blocking&lt;/code&gt; entries); (2)
add a &lt;strong&gt;transitional&lt;/strong&gt; dossier content-parity test proving software-dev’s resolved
required-artifact set is unchanged across the swap — this scaffold is &lt;strong&gt;deleted in
this lane’s final commit&lt;/strong&gt;, not merged; (3) build the &lt;code dir=&quot;auto&quot;&gt;ConfigResult&lt;/code&gt; →
&lt;code dir=&quot;auto&quot;&gt;ExpectedArtifactManifest&lt;/code&gt; adapter (&lt;code dir=&quot;auto&quot;&gt;repository.get_expected_artifacts&lt;/code&gt; returns
&lt;code dir=&quot;auto&quot;&gt;ConfigResult&lt;/code&gt;, not the model the six consumer sites expect, and there is no
&lt;code dir=&quot;auto&quot;&gt;from_dict&lt;/code&gt; today) with cache preservation; (4) flip the reader; (5) delete the
&lt;code dir=&quot;auto&quot;&gt;specify_cli&lt;/code&gt; copies. Never delete before the parity scaffold is green; delete the
scaffold once the swap is proven. The enduring dossier-behaviour assertions live
as doctrine-module + integration tests against the now-canonical doctrine tree. If
reconciliation reveals deep drift, only the final flip defers to slice 2.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The mission-instance addendum layer is specified but not built&lt;/strong&gt; (a future
&lt;code dir=&quot;auto&quot;&gt;meta.json governance_addendum&lt;/code&gt; as the highest field-merge layer) — deferred,
YAGNI.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The dual-tree split is on the deprecation path, not entrenched.&lt;/strong&gt; No new
content is added to &lt;code dir=&quot;auto&quot;&gt;specify_cli/missions/&lt;/code&gt;; and critically &lt;strong&gt;no surviving
“keep in sync” or parity ratchet is introduced&lt;/strong&gt; — a permanent parity gate would
entrench the very split this swap removes. Parity is proven transitionally and the
scaffold is deleted; the copies are removed, leaving one source.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Testing posture (operator-mandated).&lt;/strong&gt; Parity/snapshot tests are transitional
migration scaffolds — added at the start of a swap, deleted before merge. Enduring
tests verify &lt;strong&gt;behaviour&lt;/strong&gt; and shift down to &lt;strong&gt;doctrine-module unit tests +
integration-level checks&lt;/strong&gt; (e.g. “a documentation mission resolves documentation
governance and gates, not software-dev doctrine”), not byte pins on the removed
path. No code — wrapper, shim, or preserved signature — is kept solely to avoid
updating tests; the suite is expected to change substantially and is updated to
the new behaviour.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3 id=&quot;work-package-spine-as-refined-by-the-second-opinion-review&quot;&gt;Work-package spine (as refined by the second-opinion review)&lt;/h3&gt;
&lt;p&gt;Four lanes; ~10–12 WPs (Q3 adds the steps lane). Lane A: &lt;strong&gt;WP-TIDY&lt;/strong&gt; (remove the
inert &lt;code dir=&quot;auto&quot;&gt;governance_refs&lt;/code&gt; + dangling refs; behaviour-preserving; first). Lane B
(critical path): &lt;strong&gt;WP-CANON&lt;/strong&gt; (single mission-type canonicalizer) → &lt;strong&gt;WP-SEAM&lt;/strong&gt;
(resolver + ordered &lt;code dir=&quot;auto&quot;&gt;ResolvedGovernance&lt;/code&gt; + &lt;code dir=&quot;auto&quot;&gt;MissionType&lt;/code&gt;-artefact-sourced bundle +
Surface B rewire; a &lt;em&gt;transitional&lt;/em&gt; byte-parity scaffold guards the swap and is
deleted in this lane’s final commit) → &lt;strong&gt;WP-LEAK&lt;/strong&gt; (rewire the live Surface A
bundle + delete the dead pair + split &lt;code dir=&quot;auto&quot;&gt;template_set&lt;/code&gt;; red-first behavioural test
first) → &lt;strong&gt;WP-OVERRIDE&lt;/strong&gt; (per-type override ridden through the overlay stack: &lt;code dir=&quot;auto&quot;&gt;id&lt;/code&gt;
on &lt;code dir=&quot;auto&quot;&gt;MissionTypeProfile&lt;/code&gt; + &lt;code dir=&quot;auto&quot;&gt;BaseDoctrineRepository&lt;/code&gt; subclass + precedence/collision
test). Lane C (parallel authoring): &lt;strong&gt;WP-CONTENT-DOC / -RESEARCH / -PLAN&lt;/strong&gt; — split
×3 because “author 3 sets” hides &lt;strong&gt;6–8 net-new DRG-resolvable artifacts&lt;/strong&gt;
(Divio-type, plain-language, accessibility, publication, freshness-SLA styleguides;
citation discipline; and the already-referenced-but-missing &lt;code dir=&quot;auto&quot;&gt;spike-timebox-policy&lt;/code&gt;
procedure). Lane D (detachable, non-blocking): &lt;strong&gt;WP-GATES-RECONCILE&lt;/strong&gt; →
&lt;strong&gt;WP-GATES-MIGRATE&lt;/strong&gt; (gates/dossier swap; transitional dossier-parity scaffold,
deleted at the end of the lane) and &lt;strong&gt;WP-STEPS-MIGRATE&lt;/strong&gt; (route step-contract
resolution through the artefact bundle and off any &lt;code dir=&quot;auto&quot;&gt;specify_cli&lt;/code&gt; path; same
transitional-parity-then-delete discipline). Join: &lt;strong&gt;WP-ENFORCE&lt;/strong&gt; — the &lt;em&gt;enduring&lt;/em&gt;
behavioural non-leakage/non-vacuity assertions, authored as &lt;strong&gt;doctrine-module +
integration tests&lt;/strong&gt;, not a surviving parity ratchet. Transitional parity scaffolds
are added at each swap’s start and deleted before merge; enduring tests verify
behaviour and move down to the doctrine module + integration level. Tests are a
first-class, per-WP change surface — and the suite is expected to change
substantially rather than be shielded by compat shims.&lt;/p&gt;
&lt;h3 id=&quot;consequences&quot;&gt;Consequences&lt;/h3&gt;
&lt;h4 id=&quot;positive&quot;&gt;Positive&lt;/h4&gt;
&lt;ul&gt;
&lt;li&gt;The software-dev-default leak becomes impossible by construction on the
action-doctrine path in slice 1, and on the dossier path once &lt;code dir=&quot;auto&quot;&gt;WP-CANON&lt;/code&gt; +
the gated migration land.&lt;/li&gt;
&lt;li&gt;Governance collapses from three surfaces + a dead field to two URN-normalized
grains resolved through one function, overridable via the established overlay
stack.&lt;/li&gt;
&lt;li&gt;After slice 1 the core is measurably &lt;strong&gt;less&lt;/strong&gt; dependent on
&lt;code dir=&quot;auto&quot;&gt;specify_cli/missions/&lt;/code&gt; (the dossier gate reader is migrated behind a parity
gate and the derived &lt;code dir=&quot;auto&quot;&gt;expected-artifacts.yaml&lt;/code&gt; copies are deleted), and a proven
&lt;code dir=&quot;auto&quot;&gt;doctrine → charter → core&lt;/code&gt; path exists for later slices to widen to templates
and gates by filling a reserved slot and repointing one reader each.&lt;/li&gt;
&lt;li&gt;&lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt; loses its special status with no behavioural change.&lt;/li&gt;
&lt;/ul&gt;
&lt;h4 id=&quot;negative&quot;&gt;Negative&lt;/h4&gt;
&lt;ul&gt;
&lt;li&gt;The rerouting risks reordering/de-duping software-dev directives; because the
existing content-contract suite is substring-based, the swap is guarded by
&lt;strong&gt;transitional&lt;/strong&gt; byte- and dossier-parity scaffolds (deleted before merge), with
enduring behaviour verified by doctrine-module + integration tests — real test
work, and a deliberate rewrite of the suite rather than compat-shielding it.&lt;/li&gt;
&lt;li&gt;The per-type override and the dossier migration each carry named adapter cost
(overlay &lt;code dir=&quot;auto&quot;&gt;id&lt;/code&gt; handling; &lt;code dir=&quot;auto&quot;&gt;ConfigResult&lt;/code&gt; → &lt;code dir=&quot;auto&quot;&gt;ExpectedArtifactManifest&lt;/code&gt;); neither is
the “free repoint” a naive reading would assume.&lt;/li&gt;
&lt;li&gt;Slice 1 populates governance, gates, and steps; the &lt;code dir=&quot;auto&quot;&gt;template_set&lt;/code&gt; slot is
declared on the artefact but populated in a later slice.&lt;/li&gt;
&lt;/ul&gt;
&lt;h4 id=&quot;neutral&quot;&gt;Neutral&lt;/h4&gt;
&lt;ul&gt;
&lt;li&gt;This ADR does not migrate templates or the remaining &lt;code dir=&quot;auto&quot;&gt;specify_cli/missions/&lt;/code&gt;
readers (enumeration, &lt;code dir=&quot;auto&quot;&gt;mission-runtime&lt;/code&gt;, the copy step), and does not delete the
tree; those are explicit later slices of the retirement epic.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3 id=&quot;confirmation&quot;&gt;Confirmation&lt;/h3&gt;
&lt;p&gt;Two kinds of check confirm the mission — and only the first kind survives it.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Enduring (behavioural, doctrine-module + integration):&lt;/strong&gt; (1) each non-software
mission’s &lt;strong&gt;unioned (type ⊕ action)&lt;/strong&gt; resolved set is disjoint from a curated,
URN-normalized software-dev-only denylist, with a &lt;strong&gt;non-vacuity twin&lt;/strong&gt; proving
software-dev &lt;em&gt;does&lt;/em&gt; resolve that set (exercised through a &lt;em&gt;shared&lt;/em&gt; action name so it
cannot pass vacuously); (2) an &lt;em&gt;unknown&lt;/em&gt; &lt;code dir=&quot;auto&quot;&gt;mission_type&lt;/code&gt; hard-fails on every
resolution path while a &lt;em&gt;known type with an empty grain&lt;/em&gt; (e.g. &lt;code dir=&quot;auto&quot;&gt;plan&lt;/code&gt;) resolves
empty without error; (3) the doctrine &lt;code dir=&quot;auto&quot;&gt;MissionType&lt;/code&gt; artefact is the sole source of
type/steps/gates/governance and every authored governance id resolves in the DRG;
(4) an integration check exercises a real documentation/research/plan mission and
observes domain-appropriate governance and gates, no software-dev doctrine.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Transitional (parity scaffolds, deleted before merge):&lt;/strong&gt; software-dev’s rendered
governance and its resolved dossier required-artifact set are proven unchanged
across each swap (the &lt;code dir=&quot;auto&quot;&gt;lint-report&lt;/code&gt; entries reconciled upward first); these scaffolds
verify the swap is user-invisible and are then removed, so no parity ratchet
survives to re-entrench the split. The mission is &lt;em&gt;not&lt;/em&gt; confirmed by keeping the old
substring content-suite frozen; that suite is rewritten to the new behaviour.&lt;/p&gt;
&lt;h3 id=&quot;decisions-resolved-with-the-operator-formerly-open&quot;&gt;Decisions resolved with the operator (formerly open)&lt;/h3&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Q1 — Governance declaration shape:&lt;/strong&gt; the &lt;code dir=&quot;auto&quot;&gt;MissionType&lt;/code&gt; artefact &lt;strong&gt;references a
sibling &lt;code dir=&quot;auto&quot;&gt;governance-profile.yaml&lt;/code&gt;&lt;/strong&gt; for type-grain governance (the live, schema’d,
hard-failing surface the resolver already reads), unioned with the action-grain
&lt;code dir=&quot;auto&quot;&gt;actions/&amp;#x3C;action&gt;/index.yaml&lt;/code&gt;. governance-profile.yaml is kept, not absorbed into
&lt;code dir=&quot;auto&quot;&gt;mission_types/&amp;#x3C;type&gt;.yaml&lt;/code&gt;. Lower churn; the override rides the profile.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Q2 — Overlay adapter form:&lt;/strong&gt; &lt;strong&gt;ride the &lt;code dir=&quot;auto&quot;&gt;doctrine/base.py&lt;/code&gt; overlay stack.&lt;/strong&gt;
&lt;code dir=&quot;auto&quot;&gt;WP-OVERRIDE&lt;/code&gt; adds an &lt;code dir=&quot;auto&quot;&gt;id&lt;/code&gt; to &lt;code dir=&quot;auto&quot;&gt;MissionTypeProfile&lt;/code&gt; + a &lt;code dir=&quot;auto&quot;&gt;BaseDoctrineRepository&lt;/code&gt;
subclass so per-type overrides get builtin → org → project ordering and collision
warnings from the shared loader (and #832 org-layer support) rather than a bespoke
second merge site.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Q3 — Artefact depth in slice 1:&lt;/strong&gt; &lt;strong&gt;governance + gates + steps.&lt;/strong&gt; Step-contract
resolution routes through the artefact bundle in slice 1 (&lt;code dir=&quot;auto&quot;&gt;WP-STEPS-MIGRATE&lt;/code&gt;);
only &lt;code dir=&quot;auto&quot;&gt;template_set&lt;/code&gt; (templates) defers to a later slice.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Q4 — ADR meta:&lt;/strong&gt; &lt;strong&gt;Accepted, epic altitude&lt;/strong&gt; (this record; one ADR governs the
retirement arc, with #883 as slice 1).&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Genuinely still open (settled inside the named WPs, not architecture-level): the
exact denylist membership for the non-leakage test, and the precise per-entry
degrade behaviour for the mission-less &lt;code dir=&quot;auto&quot;&gt;build_charter_context&lt;/code&gt; callers.&lt;/p&gt;
&lt;h2 id=&quot;pros-and-cons-of-the-options&quot;&gt;Pros and Cons of the Options&lt;/h2&gt;
&lt;h3 id=&quot;option-1--populate-governance_refs&quot;&gt;Option 1 — Populate &lt;code dir=&quot;auto&quot;&gt;governance_refs&lt;/code&gt;&lt;/h3&gt;
&lt;p&gt;&lt;strong&gt;Pros:&lt;/strong&gt; matches the field’s apparent intent; smallest edit.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Cons:&lt;/strong&gt; inert (no runtime reader); cannot carry the tactics/styleguides/
procedures/paradigms non-software governance is made of; changes &lt;code dir=&quot;auto&quot;&gt;graph.yaml&lt;/code&gt; by
nothing, so the freshness gate gives false assurance. Leaves the leak and the
split untouched.&lt;/p&gt;
&lt;h3 id=&quot;option-2--fill-governance-profileyaml-only&quot;&gt;Option 2 — Fill &lt;code dir=&quot;auto&quot;&gt;governance-profile.yaml&lt;/code&gt; only&lt;/h3&gt;
&lt;p&gt;&lt;strong&gt;Pros:&lt;/strong&gt; ships non-software content on the live prompt path with no new plumbing.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Cons:&lt;/strong&gt; leaves the action-scoped and dossier-path leaks in place — the headline
acceptance criterion stays unmet — and does nothing toward the north star.&lt;/p&gt;
&lt;h3 id=&quot;option-3--unified-doctrine--charter--core-seam-chosen&quot;&gt;Option 3 — Unified &lt;code dir=&quot;auto&quot;&gt;doctrine → charter → core&lt;/code&gt; seam (CHOSEN)&lt;/h3&gt;
&lt;p&gt;&lt;strong&gt;Pros:&lt;/strong&gt; closes the leak on the action path (and the dossier path via &lt;code dir=&quot;auto&quot;&gt;WP-CANON&lt;/code&gt; +
the gated migration); collapses three surfaces + a dead field to two
URN-normalized grains + one resolver; reuses the overlay stack for layering
(adapter cost named, not hidden); demotes &lt;code dir=&quot;auto&quot;&gt;software-dev&lt;/code&gt; to a peer with no
behavioural change; leaves a proven, widening seam for later slices.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Cons:&lt;/strong&gt; larger than a content-only change; centralises resolution (mitigated by
a new ordered byte-snapshot gate + dossier parity gate); names real adapter work
on the override and dossier surfaces; forward-shapes reserved bundle slots.&lt;/p&gt;
&lt;h3 id=&quot;option-4--full-dual-tree-merge-now&quot;&gt;Option 4 — Full dual-tree merge now&lt;/h3&gt;
&lt;p&gt;&lt;strong&gt;Pros:&lt;/strong&gt; reaches the end-state in one mission.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Cons:&lt;/strong&gt; unbounded blast radius across every &lt;code dir=&quot;auto&quot;&gt;specify_cli/missions/&lt;/code&gt; reader in a
single change; high regression risk against software-dev; violates the slice
discipline that lets each reader migration be independently verified.&lt;/p&gt;
&lt;h2 id=&quot;more-information&quot;&gt;More Information&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Technical story: &lt;a href=&quot;https://github.com/Priivacy-ai/spec-kitty/issues/883&quot;&gt;#883&lt;/a&gt;
(slice 1) under epics &lt;a href=&quot;https://github.com/Priivacy-ai/spec-kitty/issues/461&quot;&gt;#461&lt;/a&gt;,
&lt;a href=&quot;https://github.com/Priivacy-ai/spec-kitty/issues/901&quot;&gt;#901&lt;/a&gt;; related
&lt;a href=&quot;https://github.com/Priivacy-ai/spec-kitty/issues/832&quot;&gt;#832&lt;/a&gt;,
&lt;a href=&quot;https://github.com/Priivacy-ai/spec-kitty/pull/2628&quot;&gt;#2628&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;This ADR was refined by an adversarial four-lens second-opinion review
(test-strategy, architecture-refutation, implementer-reality, slice-discipline);
the corrections — dead-code leak anchors, the cross-boundary &lt;code dir=&quot;auto&quot;&gt;mission.py:575&lt;/code&gt;
default, the non-free overlay adapter, the substring-not-byte content suite, the
drifted expected-artifacts trees, and the &lt;code dir=&quot;auto&quot;&gt;plan&lt;/code&gt; empty-grain policy — are folded
into the decisions above.&lt;/li&gt;
&lt;li&gt;Governing layer-merge contract reused:
&lt;a href=&quot;/adr/3.x/2026-05-16-1-doctrine-layer-merge-semantics/&quot;&gt;2026-05-16-1 — Doctrine-Layer Merge Semantics&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;Adjacent resolution-seam precedent:
&lt;a href=&quot;/adr/3.x/2026-07-08-1-mission-resolver-port/&quot;&gt;2026-07-08-1 — MissionResolver Port&lt;/a&gt;,
&lt;a href=&quot;/adr/3.x/2026-06-26-1-single-authority-seam-and-call-site-gate/&quot;&gt;2026-06-26-1 — Single-Authority Seam + Call-Site Gate&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;Load-bearing anchors: seam &lt;code dir=&quot;auto&quot;&gt;charter/mission_type_profiles.py:175,266,325,392,440&lt;/code&gt;;
live consumers &lt;code dir=&quot;auto&quot;&gt;runtime/next/prompt_builder.py:346&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;charter/context.py:252,865,3254&lt;/code&gt;;
dead pair to delete &lt;code dir=&quot;auto&quot;&gt;charter/context.py:1451,1500&lt;/code&gt;; canonicalizer targets
&lt;code dir=&quot;auto&quot;&gt;charter/mission_type_profiles.py:380&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;specify_cli/mission.py:575&lt;/code&gt;; overlay stack
&lt;code dir=&quot;auto&quot;&gt;doctrine/base.py:213,249,289-311&lt;/code&gt;; FSM reads &lt;code dir=&quot;auto&quot;&gt;decision.py:601&lt;/code&gt;,
&lt;code dir=&quot;auto&quot;&gt;runtime_bridge_composition.py:186&lt;/code&gt;; retirement targets dossier
&lt;code dir=&quot;auto&quot;&gt;dossier/manifest.py:163-193&lt;/code&gt; → dead doctrine reader
&lt;code dir=&quot;auto&quot;&gt;doctrine/missions/repository.py:304&lt;/code&gt;, copy step &lt;code dir=&quot;auto&quot;&gt;template/manager.py:88&lt;/code&gt;,
hardcoded software-dev &lt;code dir=&quot;auto&quot;&gt;mission.py:466,469&lt;/code&gt;; retired field &lt;code dir=&quot;auto&quot;&gt;models.py:186&lt;/code&gt;,
&lt;code dir=&quot;auto&quot;&gt;mission_type.py:1486,1505&lt;/code&gt;; dangling refs &lt;code dir=&quot;auto&quot;&gt;software-dev.yaml:11-12&lt;/code&gt;;
content suite (substring, not byte) &lt;code dir=&quot;auto&quot;&gt;tests/specify_cli/next/test_wp_prompt_governance_contract.py&lt;/code&gt;.&lt;/li&gt;
&lt;/ul&gt;</content>
</entry>
