<?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/2026-09-20-1-issue-matrix-not-applicable-verdict/">
  <id>https://docs.spec-kitty.ai/adr/3.x/2026-09-20-1-issue-matrix-not-applicable-verdict/</id>
  <title>ADR: `not-applicable` issue-matrix verdict — non-gating, terminal, classifier-narrowed approval gate</title>
  <updated>2026-09-20T00:00:00.000Z</updated>
  <author><name>Spec Kitty</name></author>
  <link rel="alternate" type="text/html" href="https://docs.spec-kitty.ai/adr/3.x/2026-09-20-1-issue-matrix-not-applicable-verdict/"/>
  <summary>Adds a non-gating, terminal not-applicable verdict and narrows the approval blocker to gate only implementation_target references, superseding the never-shipped WP09 Gate-4 intent.</summary>
  <content type="html">&lt;h2 id=&quot;context-and-problem-statement&quot;&gt;Context and Problem Statement&lt;/h2&gt;
&lt;p&gt;Mission &lt;code dir=&quot;auto&quot;&gt;move-task-approval-ergonomics-01M302R0&lt;/code&gt; (#3469) WP02, built on WP01’s reference
classifier (&lt;code dir=&quot;auto&quot;&gt;specify_cli.tasks.issue_reference_discovery&lt;/code&gt;). WP01 landed a pure,
deterministic classifier (&lt;code dir=&quot;auto&quot;&gt;classify_occurrences&lt;/code&gt; / &lt;code dir=&quot;auto&quot;&gt;GatingClass&lt;/code&gt;) that labels every
discovered &lt;code dir=&quot;auto&quot;&gt;#NNNN&lt;/code&gt; reference as &lt;code dir=&quot;auto&quot;&gt;implementation_target&lt;/code&gt; (gating), &lt;code dir=&quot;auto&quot;&gt;context_only&lt;/code&gt;, or
&lt;code dir=&quot;auto&quot;&gt;pr_or_commit_ref&lt;/code&gt; (both non-gating), plus the single shared predicate &lt;code dir=&quot;auto&quot;&gt;is_gating()&lt;/code&gt; /
&lt;code dir=&quot;auto&quot;&gt;gating_issue_numbers()&lt;/code&gt; every consumer must derive “referenced-but-missing” from
(FR-012). WP02 is the first consumer to wire that classifier into a live enforcement
site: the &lt;code dir=&quot;auto&quot;&gt;move-task&lt;/code&gt;/approval blocker in
&lt;code dir=&quot;auto&quot;&gt;src/specify_cli/cli/commands/agent/tasks_parsing_validation.py&lt;/code&gt;.&lt;/p&gt;
&lt;p&gt;Two problems motivated this change:&lt;/p&gt;
&lt;ol&gt;
&lt;li&gt;&lt;strong&gt;No truthful non-gating verdict.&lt;/strong&gt; &lt;code dir=&quot;auto&quot;&gt;IssueMatrixVerdict&lt;/code&gt; (&lt;code dir=&quot;auto&quot;&gt;src/specify_cli/cli/ commands/review/_issue_matrix.py&lt;/code&gt;) offered four values: &lt;code dir=&quot;auto&quot;&gt;fixed&lt;/code&gt;,
&lt;code dir=&quot;auto&quot;&gt;verified-already-fixed&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;deferred-with-followup&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;in-mission&lt;/code&gt;. None of them let an
operator honestly record “this reference is cited for context, or is a PR/commit
reference — the mission owes it no work.” An operator resolving a context-only row
was forced to pick a verdict that asserted work that never happened (&lt;code dir=&quot;auto&quot;&gt;fixed&lt;/code&gt;) or an
unintended follow-up promise (&lt;code dir=&quot;auto&quot;&gt;deferred-with-followup&lt;/code&gt;).&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The approval blocker gated on every discovered reference, not just gating ones.&lt;/strong&gt;
&lt;code dir=&quot;auto&quot;&gt;_issue_matrix_approval_blocker&lt;/code&gt; (and its internal &lt;code dir=&quot;auto&quot;&gt;_issue_matrix_evaluation&lt;/code&gt; helper)
computed &lt;code dir=&quot;auto&quot;&gt;referenced_issues = {f&quot;#{ref.number}&quot; for ref in refs}&lt;/code&gt; — every reference
&lt;code dir=&quot;auto&quot;&gt;discover_issue_references&lt;/code&gt; found, with no consultation of WP01’s classification at
all. A mission that referenced only &lt;code dir=&quot;auto&quot;&gt;context_only&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;pr_or_commit_ref&lt;/code&gt; issues (e.g. a
&lt;code dir=&quot;auto&quot;&gt;baseline-red until #200 lands&lt;/code&gt; note, or a &lt;code dir=&quot;auto&quot;&gt;PR #300&lt;/code&gt; citation) was still forced to
scaffold an &lt;code dir=&quot;auto&quot;&gt;issue-matrix.json&lt;/code&gt; and fill a row for every one of them before &lt;code dir=&quot;auto&quot;&gt;approved&lt;/code&gt;
— the classifier existed but nothing consumed it yet.&lt;/li&gt;
&lt;/ol&gt;
&lt;h3 id=&quot;a-previously-documented-never-implemented-alternative&quot;&gt;A previously-documented, never-implemented alternative&lt;/h3&gt;
&lt;p&gt;&lt;code dir=&quot;auto&quot;&gt;src/specify_cli/policy/merge_gates.py::_evaluate_issue_matrix_completeness_gate&lt;/code&gt; already
carried this comment before this ADR:&lt;/p&gt;
&lt;blockquote&gt;
&lt;p&gt;“Fail-closed only when references exist: zero discovered references is a PASS (nothing
to enforce). WP09 owns the formal &lt;code dir=&quot;auto&quot;&gt;not_applicable&lt;/code&gt; Gate-4 verdict for the post-merge
review surface; this merge gate’s zero-reference branch is intentionally the simpler
“nothing to check” case, not a re-definition of &lt;code dir=&quot;auto&quot;&gt;not_applicable&lt;/code&gt;.”&lt;/p&gt;
&lt;/blockquote&gt;
&lt;p&gt;No “WP09” mission ever implemented a &lt;code dir=&quot;auto&quot;&gt;not_applicable&lt;/code&gt; Gate-4 verdict anywhere in this
repository — grepping the tree turns up only this one comment, describing an intent that
was never realized. Left standing, that comment invited a second, independently-invented
&lt;code dir=&quot;auto&quot;&gt;not_applicable&lt;/code&gt; definition on the merge-gate surface whenever someone eventually acted
on it, which would have diverged from whatever vocabulary landed first on the
&lt;code dir=&quot;auto&quot;&gt;issue-verdict&lt;/code&gt;/approval-blocker surface (exactly the two-definitions-diverge failure
mode FR-010 calls out). &lt;strong&gt;This ADR is that first landing, and it explicitly supersedes
the WP09 comment’s intent&lt;/strong&gt;: there is now exactly one &lt;code dir=&quot;auto&quot;&gt;not_applicable&lt;/code&gt;-shaped verdict —
&lt;code dir=&quot;auto&quot;&gt;IssueMatrixVerdict.NOT_APPLICABLE&lt;/code&gt; — and any future work on &lt;code dir=&quot;auto&quot;&gt;merge_gates.py&lt;/code&gt;’s
&lt;code dir=&quot;auto&quot;&gt;_evaluate_issue_matrix_completeness_gate&lt;/code&gt; (mission &lt;code dir=&quot;auto&quot;&gt;move-task-approval-ergonomics- 01M302R0&lt;/code&gt; WP03, cross-site adoption) must consume this same value and the same
classifier predicate (&lt;code dir=&quot;auto&quot;&gt;is_gating&lt;/code&gt;), not invent a second Gate-4-specific enum or string
literal. The comment in &lt;code dir=&quot;auto&quot;&gt;merge_gates.py&lt;/code&gt; is left in place (WP02 does not own that file)
but is superseded in intent as of this ADR; WP03 updates or removes it when it adopts the
classifier.&lt;/p&gt;
&lt;h2 id=&quot;decision&quot;&gt;Decision&lt;/h2&gt;
&lt;h3 id=&quot;d1--add-issuematrixverdictnot_applicable--not-applicable-additive&quot;&gt;D1 — Add &lt;code dir=&quot;auto&quot;&gt;IssueMatrixVerdict.NOT_APPLICABLE = &quot;not-applicable&quot;&lt;/code&gt; (additive)&lt;/h3&gt;
&lt;p&gt;Appended as a fifth member of the &lt;code dir=&quot;auto&quot;&gt;StrEnum&lt;/code&gt;, after &lt;code dir=&quot;auto&quot;&gt;IN_MISSION&lt;/code&gt;. No existing member is
renamed, reordered, or removed. Every consumer that parses a verdict does so via
&lt;code dir=&quot;auto&quot;&gt;IssueMatrixVerdict(value)&lt;/code&gt; (a plain &lt;code dir=&quot;auto&quot;&gt;try&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;except ValueError&lt;/code&gt;) rather than an exhaustive
&lt;code dir=&quot;auto&quot;&gt;match&lt;/code&gt;, so the four legacy values continue to parse and validate byte-for-byte
unchanged — an existing &lt;code dir=&quot;auto&quot;&gt;issue-matrix.json&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;issue-matrix.md&lt;/code&gt; needs no migration
(NFR-001, SC-006).&lt;/p&gt;
&lt;h3 id=&quot;d2--not-applicable-is-non-gating-at-approved-and-terminal-at-donemerge&quot;&gt;D2 — &lt;code dir=&quot;auto&quot;&gt;not-applicable&lt;/code&gt; is non-gating at &lt;code dir=&quot;auto&quot;&gt;approved&lt;/code&gt; and terminal at &lt;code dir=&quot;auto&quot;&gt;done&lt;/code&gt;/merge&lt;/h3&gt;
&lt;p&gt;Unlike &lt;code dir=&quot;auto&quot;&gt;in-mission&lt;/code&gt; (non-terminal: accepted at &lt;code dir=&quot;auto&quot;&gt;approved&lt;/code&gt;, rejected at &lt;code dir=&quot;auto&quot;&gt;done&lt;/code&gt; via the
&lt;code dir=&quot;auto&quot;&gt;_issue_matrix_approval_blocker&lt;/code&gt;’s &lt;code dir=&quot;auto&quot;&gt;unresolved_in_mission&lt;/code&gt; special-case list),
&lt;code dir=&quot;auto&quot;&gt;not-applicable&lt;/code&gt; requires &lt;strong&gt;no special-case code at either transition&lt;/strong&gt;. Once a row
exists with verdict &lt;code dir=&quot;auto&quot;&gt;not-applicable&lt;/code&gt;:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;It is a &lt;em&gt;valid&lt;/em&gt; &lt;code dir=&quot;auto&quot;&gt;IssueMatrixVerdict&lt;/code&gt; member, so &lt;code dir=&quot;auto&quot;&gt;validate_issue_matrix&lt;/code&gt;/
&lt;code dir=&quot;auto&quot;&gt;diagnose_structured_issue_matrix&lt;/code&gt; accept the row (assuming a non-empty
&lt;code dir=&quot;auto&quot;&gt;evidence_ref&lt;/code&gt; — the one universal row rule) at both &lt;code dir=&quot;auto&quot;&gt;approved&lt;/code&gt; and &lt;code dir=&quot;auto&quot;&gt;done&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;It is never added to &lt;code dir=&quot;auto&quot;&gt;unresolved_in_mission&lt;/code&gt; (that list only ever collects rows whose
verdict &lt;code dir=&quot;auto&quot;&gt;is IssueMatrixVerdict.IN_MISSION&lt;/code&gt;).&lt;/li&gt;
&lt;li&gt;It therefore never re-blocks at merge — it is terminal by the &lt;em&gt;absence&lt;/em&gt; of a rejection
rule, not by an added exemption. This is deliberate: there is no future state in which
this mission “resolves” a reference it never owed work on, so there is nothing for a
&lt;code dir=&quot;auto&quot;&gt;done&lt;/code&gt;-time check to wait for.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3 id=&quot;d3--the-approval-blocker-gates-only-wp01-classified-implementation_target-references&quot;&gt;D3 — The approval blocker gates only WP01-classified &lt;code dir=&quot;auto&quot;&gt;implementation_target&lt;/code&gt; references&lt;/h3&gt;
&lt;p&gt;&lt;code dir=&quot;auto&quot;&gt;_issue_matrix_approval_blocker&lt;/code&gt; and &lt;code dir=&quot;auto&quot;&gt;_issue_matrix_evaluation&lt;/code&gt;
(&lt;code dir=&quot;auto&quot;&gt;tasks_parsing_validation.py&lt;/code&gt;) now import &lt;code dir=&quot;auto&quot;&gt;is_gating&lt;/code&gt; from
&lt;code dir=&quot;auto&quot;&gt;specify_cli.tasks.issue_reference_discovery&lt;/code&gt; and filter every discovered reference
through it before computing &lt;code dir=&quot;auto&quot;&gt;referenced_issues&lt;/code&gt; / &lt;code dir=&quot;auto&quot;&gt;missing_issues&lt;/code&gt; /
&lt;code dir=&quot;auto&quot;&gt;unresolved_in_mission&lt;/code&gt;, and before deciding whether an issue-matrix artifact is
required at all:&lt;/p&gt;
&lt;div class=&quot;expressive-code&quot;&gt;&lt;figure class=&quot;frame not-content&quot;&gt;&lt;figcaption class=&quot;header&quot;&gt;&lt;/figcaption&gt;&lt;pre data-language=&quot;python&quot;&gt;&lt;code&gt;&lt;div class=&quot;ec-line&quot;&gt;&lt;div class=&quot;code&quot;&gt;&lt;span style=&quot;--0:#D6DEEB;--1:#403F53&quot;&gt;gating_refs &lt;/span&gt;&lt;span style=&quot;--0:#C792EA;--1:#8844AE&quot;&gt;=&lt;/span&gt;&lt;span style=&quot;--0:#D6DEEB;--1:#403F53&quot;&gt; &lt;/span&gt;&lt;span style=&quot;--0:#D9F5DD;--1:#111111&quot;&gt;[&lt;/span&gt;&lt;span style=&quot;--0:#D6DEEB;--1:#403F53&quot;&gt;ref &lt;/span&gt;&lt;span style=&quot;--0:#C792EA;--1:#8844AE&quot;&gt;for&lt;/span&gt;&lt;span style=&quot;--0:#D6DEEB;--1:#403F53&quot;&gt; ref &lt;/span&gt;&lt;span style=&quot;--0:#C792EA;--1:#8844AE&quot;&gt;in&lt;/span&gt;&lt;span style=&quot;--0:#D6DEEB;--1:#403F53&quot;&gt; refs &lt;/span&gt;&lt;span style=&quot;--0:#C792EA;--1:#8844AE&quot;&gt;if&lt;/span&gt;&lt;span style=&quot;--0:#D6DEEB;--1:#403F53&quot;&gt; &lt;/span&gt;&lt;span style=&quot;--0:#B2CCD6;--1:#096E72&quot;&gt;is_gating&lt;/span&gt;&lt;span style=&quot;--0:#D6DEEB;--1:#403F53&quot;&gt;(&lt;/span&gt;&lt;span style=&quot;--0:#82AAFF;--1:#3B61B0&quot;&gt;ref&lt;/span&gt;&lt;span style=&quot;--0:#D6DEEB;--1:#403F53&quot;&gt;)&lt;/span&gt;&lt;span style=&quot;--0:#D9F5DD;--1:#111111&quot;&gt;]&lt;/span&gt;&lt;/div&gt;&lt;/div&gt;&lt;div class=&quot;ec-line&quot;&gt;&lt;div class=&quot;code&quot;&gt;&lt;span style=&quot;--0:#C792EA;--1:#8844AE&quot;&gt;if&lt;/span&gt;&lt;span style=&quot;--0:#D6DEEB;--1:#403F53&quot;&gt; &lt;/span&gt;&lt;span style=&quot;--0:#C792EA;--1:#8844AE&quot;&gt;not&lt;/span&gt;&lt;span style=&quot;--0:#D6DEEB;--1:#403F53&quot;&gt; gating_refs:&lt;/span&gt;&lt;/div&gt;&lt;/div&gt;&lt;div class=&quot;ec-line&quot;&gt;&lt;div class=&quot;code&quot;&gt;&lt;span class=&quot;indent&quot;&gt;    &lt;/span&gt;&lt;span style=&quot;--0:#C792EA;--1:#8844AE&quot;&gt;return&lt;/span&gt;&lt;span style=&quot;--0:#D6DEEB;--1:#403F53&quot;&gt; &lt;/span&gt;&lt;span style=&quot;--0:#FF6A83;--1:#A24848&quot;&gt;None&lt;/span&gt;&lt;/div&gt;&lt;/div&gt;&lt;/code&gt;&lt;/pre&gt;&lt;/figure&gt;&lt;/div&gt;
&lt;p&gt;A mission whose only references are &lt;code dir=&quot;auto&quot;&gt;context_only&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;pr_or_commit_ref&lt;/code&gt; now requires &lt;strong&gt;no
issue-matrix artifact at all&lt;/strong&gt; — not even a scaffolded one — because classification
decides row-&lt;em&gt;requirement&lt;/em&gt; (FR-013), and non-gating classifications never require a row.
This is a narrowing of what blocks approval, not a widening: every reference that
gated before this change (an unmarked bare &lt;code dir=&quot;auto&quot;&gt;#NNNN&lt;/code&gt;, the FR-011 fail-safe default) still
gates identically. Only references the classifier positively demotes (an explicit &lt;code dir=&quot;auto&quot;&gt;PR #N&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;pull&lt;/code&gt; token or a &lt;code dir=&quot;auto&quot;&gt;Follow-up:&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;baseline-red&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;see #&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;parent&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;epic&lt;/code&gt; context marker)
stop requiring a row.&lt;/p&gt;
&lt;h3 id=&quot;d4--lever-ssot-classification-governs-requirement-verdict-governs-resolution-fr-013&quot;&gt;D4 — Lever SSOT: classification governs requirement, verdict governs resolution (FR-013)&lt;/h3&gt;
&lt;p&gt;Two independent levers, pinned to distinct roles, with explicit behavior where they
appear to disagree:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Classification decides whether a row is REQUIRED.&lt;/strong&gt; Only &lt;code dir=&quot;auto&quot;&gt;implementation_target&lt;/code&gt;
references ever force a row to exist.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Verdict decides whether an existing row is RESOLVED.&lt;/strong&gt; An operator may record &lt;em&gt;any&lt;/em&gt;
verdict — including &lt;code dir=&quot;auto&quot;&gt;not-applicable&lt;/code&gt; — on &lt;em&gt;any&lt;/em&gt; row, regardless of what the classifier
said about the underlying reference. Recording &lt;code dir=&quot;auto&quot;&gt;not-applicable&lt;/code&gt; on a reference the
classifier calls &lt;code dir=&quot;auto&quot;&gt;implementation_target&lt;/code&gt; is an explicit operator override (allowed);
the classifier never auto-assigns &lt;code dir=&quot;auto&quot;&gt;not-applicable&lt;/code&gt; itself — it only ever produces
&lt;code dir=&quot;auto&quot;&gt;GatingClass&lt;/code&gt; labels, never &lt;code dir=&quot;auto&quot;&gt;IssueMatrixVerdict&lt;/code&gt; values. Conversely, an
&lt;code dir=&quot;auto&quot;&gt;implementation_target&lt;/code&gt; classification requires a row to exist even if none was ever
scaffolded — a missing row is &lt;code dir=&quot;auto&quot;&gt;missing_issues&lt;/code&gt;, not silently exempted, regardless of
what verdict an operator might eventually want to put there.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3 id=&quot;d5--cli-surface---verdict-and---evidence-ref-document-both-newexisting-rules&quot;&gt;D5 — CLI surface: &lt;code dir=&quot;auto&quot;&gt;--verdict&lt;/code&gt; and &lt;code dir=&quot;auto&quot;&gt;--evidence-ref&lt;/code&gt; document both new/existing rules&lt;/h3&gt;
&lt;p&gt;&lt;code dir=&quot;auto&quot;&gt;spec-kitty agent issue-verdict --help&lt;/code&gt; now lists &lt;code dir=&quot;auto&quot;&gt;not-applicable&lt;/code&gt; alongside the four
legacy values with its non-gating/terminal meaning (NFR-004, contract C5.2), and
&lt;code dir=&quot;auto&quot;&gt;--evidence-ref --help&lt;/code&gt; documents the pre-existing &lt;code dir=&quot;auto&quot;&gt;deferred-with-followup&lt;/code&gt; evidence-
token rule (a &lt;code dir=&quot;auto&quot;&gt;#NNN&lt;/code&gt; or &lt;code dir=&quot;auto&quot;&gt;Follow-up:&lt;/code&gt; substring, previously enforced but undocumented —
FR-006) so both gating rules on the verdict surface are discoverable pre-flight rather
than discovered only after a validation failure.&lt;/p&gt;
&lt;h2 id=&quot;consequences&quot;&gt;Consequences&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Positive:&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;An operator can now record the truth on a context-only/PR-reference row instead of
a verdict that overstates work done or promises a follow-up that will never come.&lt;/li&gt;
&lt;li&gt;A mission that only ever cites non-gating references needs no issue-matrix artifact —
removing a category of unnecessary friction the classifier existed to solve but that,
before this WP, nothing consumed.&lt;/li&gt;
&lt;li&gt;&lt;code dir=&quot;auto&quot;&gt;merge_gates.py&lt;/code&gt; and &lt;code dir=&quot;auto&quot;&gt;status/doctor&lt;/code&gt; (WP03) now have exactly one &lt;code dir=&quot;auto&quot;&gt;not_applicable&lt;/code&gt;-
shaped vocabulary member and one classifier predicate to adopt, rather than reinventing
the WP09-intended Gate-4 verdict independently.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;Trade-offs / follow-ups:&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;WP03 (cross-site adoption) still owns wiring &lt;code dir=&quot;auto&quot;&gt;merge_gates.py&lt;/code&gt;’s
&lt;code dir=&quot;auto&quot;&gt;_evaluate_issue_matrix_completeness_gate&lt;/code&gt; and &lt;code dir=&quot;auto&quot;&gt;status/doctor&lt;/code&gt;’s equivalent check onto
the same &lt;code dir=&quot;auto&quot;&gt;is_gating&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;gating_issue_numbers&lt;/code&gt; predicate this ADR’s approval-blocker
change already consumes (NFR-006 cross-site consistency: a reference non-gating at
&lt;code dir=&quot;auto&quot;&gt;approved&lt;/code&gt; must be non-gating at merge and report clean in doctor). Until WP03 lands,
the merge-time gate and doctor’s divergence check still gate on every discovered
reference — a narrower gap than before WP01/WP02 (the approval blocker is now correct),
but not yet fully closed end-to-end.&lt;/li&gt;
&lt;li&gt;The &lt;code dir=&quot;auto&quot;&gt;merge_gates.py&lt;/code&gt; comment naming “WP09” is left in place (out of this WP’s owned
files) pending WP03’s pass over that module; this ADR is the authoritative record that
its &lt;code dir=&quot;auto&quot;&gt;not_applicable&lt;/code&gt; intent is superseded by &lt;code dir=&quot;auto&quot;&gt;IssueMatrixVerdict.NOT_APPLICABLE&lt;/code&gt;.&lt;/li&gt;
&lt;/ul&gt;
&lt;h2 id=&quot;alternatives-considered&quot;&gt;Alternatives Considered&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Reuse &lt;code dir=&quot;auto&quot;&gt;in-mission&lt;/code&gt; for context-only rows.&lt;/strong&gt; Rejected: &lt;code dir=&quot;auto&quot;&gt;in-mission&lt;/code&gt; asserts “a later
WP in this mission will resolve this,” which is false for a reference the mission
never owed work on, and &lt;code dir=&quot;auto&quot;&gt;in-mission&lt;/code&gt; is explicitly non-terminal (rejected at &lt;code dir=&quot;auto&quot;&gt;done&lt;/code&gt;) —
the opposite of the terminal semantics a context-only reference needs.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Skip the enum member; treat “row exists at all” as sufficient regardless of
verdict.&lt;/strong&gt; Rejected: this would silently launder an unresolved &lt;code dir=&quot;auto&quot;&gt;fixed&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;deferred- with-followup&lt;/code&gt; row that never got its real verdict recorded, and gives up the
truthfulness goal (User Story 2) entirely — the row would say nothing honest about
&lt;em&gt;why&lt;/em&gt; no further work is owed.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Gate on &lt;code dir=&quot;auto&quot;&gt;context_only&lt;/code&gt; too, only exempting &lt;code dir=&quot;auto&quot;&gt;pr_or_commit_ref&lt;/code&gt;.&lt;/strong&gt; Rejected: contradicts
WP01’s own classification contract (C1.2/C1.3), which defines both as non-gating
outcomes for the same reason (cited but not an implementation target); treating them
differently at the enforcement site would silently re-diverge from the single-source
classifier WP01 built specifically to prevent that (FR-012).&lt;/li&gt;
&lt;/ul&gt;
&lt;h2 id=&quot;references&quot;&gt;References&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Mission &lt;code dir=&quot;auto&quot;&gt;kitty-specs/move-task-approval-ergonomics-01M302R0/&lt;/code&gt;: &lt;code dir=&quot;auto&quot;&gt;spec.md&lt;/code&gt; (FR-003,
FR-004, FR-006, FR-010, FR-011, FR-012, FR-013, FR-014, NFR-001, NFR-005, NFR-006),
&lt;code dir=&quot;auto&quot;&gt;data-model.md&lt;/code&gt; (&lt;code dir=&quot;auto&quot;&gt;IssueMatrixVerdict&lt;/code&gt; value table), &lt;code dir=&quot;auto&quot;&gt;contracts/classification-and- verdict-contract.md&lt;/code&gt; (C3, C4).&lt;/li&gt;
&lt;li&gt;WP01 classifier: &lt;code dir=&quot;auto&quot;&gt;src/specify_cli/tasks/issue_reference_discovery.py&lt;/code&gt;
(&lt;code dir=&quot;auto&quot;&gt;GatingClass&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;classify_occurrences&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;is_gating&lt;/code&gt;, &lt;code dir=&quot;auto&quot;&gt;gating_issue_numbers&lt;/code&gt;).&lt;/li&gt;
&lt;li&gt;This WP’s changes: &lt;code dir=&quot;auto&quot;&gt;src/specify_cli/cli/commands/review/_issue_matrix.py&lt;/code&gt;
(&lt;code dir=&quot;auto&quot;&gt;IssueMatrixVerdict.NOT_APPLICABLE&lt;/code&gt;), &lt;code dir=&quot;auto&quot;&gt;src/specify_cli/cli/commands/agent/ issue_verdict.py&lt;/code&gt; (&lt;code dir=&quot;auto&quot;&gt;--verdict&lt;/code&gt;/&lt;code dir=&quot;auto&quot;&gt;--evidence-ref&lt;/code&gt; help), &lt;code dir=&quot;auto&quot;&gt;src/specify_cli/cli/commands/ agent/tasks_parsing_validation.py&lt;/code&gt; (&lt;code dir=&quot;auto&quot;&gt;_issue_matrix_evaluation&lt;/code&gt;,
&lt;code dir=&quot;auto&quot;&gt;_issue_matrix_approval_blocker&lt;/code&gt;).&lt;/li&gt;
&lt;li&gt;Superseded intent: &lt;code dir=&quot;auto&quot;&gt;src/specify_cli/policy/merge_gates.py:: _evaluate_issue_matrix_completeness_gate&lt;/code&gt; (the “WP09 … &lt;code dir=&quot;auto&quot;&gt;not_applicable&lt;/code&gt; Gate-4”
comment).&lt;/li&gt;
&lt;li&gt;Regression test: &lt;code dir=&quot;auto&quot;&gt;tests/specify_cli/cli/commands/test_issue_matrix_not_applicable.py&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;#4330 (unrelated, non-regressed): the JSON-first error prefix and batched
missing-row diagnostic surfacing in &lt;code dir=&quot;auto&quot;&gt;_issue_matrix_approval_blocker&lt;/code&gt; predate this ADR
and are unchanged by it.&lt;/li&gt;
&lt;/ul&gt;</content>
</entry>
