| Project alignment | "We keep mixing responsibilities, working method, plans, and what actually happened." | A.1.1; A.15; A.15.2; A.15.3; B.5.1 | F.11 when method vocabulary or work vocabulary is itself unstable; F.9 when bridge discipline is live; F.17 as a typical vocabulary-stabilizing output | the right alignment pattern is opened, or a first shared work/term form is stable enough to proceed | not when the live entry load is already comparison, boundary claim routing, or SoTA/generator scaffold design | I.2.1 gives compact-index-only posture; ToC cues stay sparse |
| Partly-said cue and language-state discovery | "Something important is there, but it is too early to publish as a settled claim, requirement, or work record." | C.2.LS; A.16; A.16.1; A.16.2; B.4.1; B.5.2.0 | endpoint claim, action, or quality patterns become candidates only after the cue is mature enough | cue preserved, language-state cue typed, or entry plurality opened without endpoint hardening | not when the claim is already stable enough for a L/A/D/E-classified boundary claim set or endpoint record | I.2.2 worked reading; lexical cues may mention "vague cue", "not yet a claim" |
| Boundary unpacking and claim decomposition | "A contract, API, protocol, SLA, acceptance, or compliance sentence mixes law, gate, duty, evidence, or action." | A.6; A.6.B; A.6.C | A.6.RSIG if first-contact recognition of the boundary description is still live; A.6.P, A.6.Q, A.6.A when relation, quality, or action wording is live | boundary claim pattern opened, or an L/A/D/E-classified atomic claim set or Claim Register is ready for the next governing FPF pattern | not when the phrase is only a partly-said cue, or when a full L/A/D/E-classified claim set already exists | I.2.3 worked reading; ToC cues should not turn API wording into generic contract authority |
| Admissible comparison, candidate-pool policy, selection, and selected-set publication | "We need comparison, a shortlist, a live pool, a call-planning distinction, or a selected set without forcing one winner too early." | A.19:0; A.17-A.19; A.19.CN; C.18; C.19; G.0; G.5 | C.11 when the entry load narrows to one local decision doctrine; C.24 when the next honest C.24 object is CallPlan or CheckpointReturn; A.19.CPM and A.19.SelectorMechanism when comparator/selector structure is live | candidate-pool policy, comparison substrate, local choice, call-plan, or selected-set publication pattern identified honestly | not when a selector mechanism or selected-set publication pattern is already settled elsewhere | I.2.4 worked reading; lexical cues may include "shortlist not winner" and "acceptable option set" |
| Generator, SoTA, or portfolio kit | "The work is to publish a reusable search, harvest, generator, selector, or portfolio scaffold, not one recommendation." | A.0; G.0; G.1; G.2; G.5 | B.5.2.1 and C.17-C.19 when creative search, novelty, or explore/exploit policy is already central; G.10 or G.11 when shipping or refresh is live | kit or scaffold entry load opened, or portfolio or set publication pattern identified | not when the entry load is only one local comparison or one one-off recommendation | I.2.5 gives compact-index-only posture unless repeated misclassification makes depth necessary |
| Same-entity rewrite, explanation, and comparative reading | "We need to restate, explain, render, repair, or compare the same claim-bearing PublicationUnit without quietly changing what it is about." | A.6.3.CR; A.6.3.RT; E.17.EFP; E.17.ID.CR | E.17.AUD.LHR and E.17.AUD.OOTD when pressured-head repair or PublicationUnit stability is live | same-entity rewrite, representation transition, explanation-facing rendering, or bounded comparative reading opened | not when the entry load is one new U.Episteme, new rule track, or independent PublicationUnit | I.2.6 worked reading; ToC cues should include "same unit, different audience" |
| Temporal claim adequacy under effort, window, and resistance | "This should speed up, slow down, recover sooner, stabilize, keep cadence, or improve throughput under a changed effort, tool-use, rollout, or policy." | C.27; C.16 when only measurement is live; A.3.3 when reusable transition law or formal model is live | B.1.4, B.1.6, C.18.1, C.19, C.22.1, C.24, C.25, C.26, C.26.3, or G.9 as the other question requires | ordinary prose, Dyn0, Dyn1 with C.16 when measurement construction or comparability is live, Dyn2TemporalClaimAdequacyCard, Dyn2TemporalClaimProfile, or a named neighboring FPF pattern relation | not when the phrase is only a speed metaphor, one state reading or snapshot, one measured rate, a service promise, a benchmark harness, or a residual QL cue without an intervention-sensitive temporal claim | I.2.7 state-to-rate-to-Dyn2 worked reading; lexical cues: speed, velocity, rhythm, cadence, throughput, recovery, braking, stabilization |
| Causal-use and counterfactual-support repair | "We want to say this caused that, this intervention would work, this policy would have prevented harm, this fairness result is causal, or this method is better on a counterfactual benchmark." | C.28; A.10; B.3; D.5; G.5; G.9 | C.16 when only a metric, score, or reading is live; C.27 when only state, rate, or intervention-sensitive temporal adequacy is live; C.26 when the phrase is only a residual quantum-like modeling cue; A.15 or A.3.2 when the question is only method, work-plan, or work-occurrence structure; A.6 when a mixed causal/deontic boundary sentence must be split | causal-use triage/card names rung, claim kind, estimand, support basis, support verdict, supported use, and unsupported use; or the wording is downgraded to association, metric, temporal, simulation-only, QL, method, work-plan, work-occurrence, or boundary-claim support | not when the sentence only records observed association, one measured metric, one process execution, one schedule, one boundary duty, or one simulation trace with no causal-use claim | I.2.8 worked reading; lexical cues: caused, would have prevented, effect, intervention, counterfactual, target trial, policy optimality, causal fairness, causal evidence, counterfactual data, method improves |