Reduced coverage playbook
When evidence is partial, keep the label attached.
Reduced coverage is an evidence state and a handoff trigger. It tells readers what TraceMap could still inspect, which proof path is incomplete, which owner role should collect the next evidence, and where the public statement must stop.
Public claim level: concept. No public conclusion without evidence. This playbook teaches labeling and owner handoff; it is not a scanner result, reducer result, runtime proof, release gate, or approval flow.
Meaning
Reduced coverage names a boundary, not a conclusion.
Partial evidence can still help reviewRule IDs, evidence tiers, coverage labels, limitations, commit context, extractor-version context, and public-safe proof links can route follow-up when they stay visible.
Fallback evidence stays labeledSyntax fallback, structure matches, config scans, project files, and public-safe gap summaries are review input until semantic or owner evidence is collected.
Missing proof is preservedUnknown, unavailable, private-only, stale, future-only, and analysis-gap labels must not be normalized into stronger wording.
Labeling
Carry the reduced label beside the sentence.
coverage labelUse the visible reduced, partial, syntax fallback, unknown, unavailable, private-only, stale, future-only, or analysis-gap label from the cited surface.evidence tierUse only Tier1Semantic, Tier2Structural, Tier3SyntaxOrTextual, or Tier4Unknown; private-only or unavailable source evidence uses Tier4Unknown with that marker.proof contextAttach rule ID or rule family, public-safe source context, extractor-version context, limitation, what cannot be concluded, next owner role, safe wording, and stop condition.Matrix
Use the row that matches the evidence gap.
| Scenario | Coverage label | Evidence tier | Evidence available | What cannot be concluded | Next owner | Safe wording | Stop condition | Proof/validation link |
|---|---|---|---|---|---|---|---|---|
| build/load failure | reduced build/load coverage | Tier3SyntaxOrTextual or Tier4Unknown unavailable |
Syntax fallback, config/project scan, public-safe gap summary, or analyzer-summary category when present. | Clean-repo claim, complete analysis, compiler-resolved conclusion, or absence-of-impact proof. | Build/tooling owner or scanner owner. | Coverage is reduced because build/load evidence is unavailable; use syntax and gap evidence as review input only. | Stop before clean-repo, complete-coverage, release-safety, or current approval wording. | Validation checks |
| syntax fallback | syntax fallback coverage | Tier3SyntaxOrTextual |
Syntax-only references, textual matches, line spans, rule family, and gap label. | Compiler-resolved symbol conclusion, call graph certainty, or complete dependency path. | Scanner owner or reviewer. | Syntax fallback found a static reference that needs semantic or owner review. | Stop before semantic, runtime, or impact conclusions. | Proof path FAQ |
| missing semantic evidence | missing semantic coverage | Tier2Structural, Tier3SyntaxOrTextual, or Tier4Unknown unavailable |
Structural or syntax evidence, unresolved symbol note, and limitation. | Tier1Semantic conclusion, compiler-resolved ownership, or complete path. |
Scanner owner or build/tooling owner. | Semantic evidence is missing, so the claim stays reduced and needs owner follow-up. | Stop before upgrading to Tier1Semantic or removing the limitation. |
Limitations and non-claims |
| unsupported framework surface | framework surface coverage gap | Tier2Structural, Tier3SyntaxOrTextual, or Tier4Unknown |
Known framework limitation, structural clue, or public-safe support note. | Complete framework coverage, route completeness, or runtime behavior. | Framework owner or scanner owner. | This framework surface is outside current public proof, so the row is a coverage gap. | Stop before complete-framework or complete-route wording. | Limitations and non-claims |
| missing generated artifact | public proof artifact unavailable | Tier1Semantic, Tier2Structural, Tier3SyntaxOrTextual, or Tier4Unknown unavailable |
Source-side evidence summary, scan-manifest family, or recorded artifact gap. | Public proof-link claim, reproducible public demo evidence, or complete artifact set. | Artifact publisher or site owner. | The public-safe artifact is unavailable, so the claim needs a replacement proof link or must stay internal. | Stop before linking, publishing, or repeating the claim publicly. | Proof path FAQ |
| private-only support | private-only coverage | Tier4Unknown private-only |
Internal evidence category, public-safe limitation, and owner note. | Public proof, customer-specific conclusion, or public demo support. | Site owner or reviewer. | Evidence may support internal follow-up, but public wording needs a public-safe summary first. | Stop before publishing private/raw details or public proof wording. | Review claim checklist |
| stale commit context | stale source context | Tier1Semantic, Tier2Structural, Tier3SyntaxOrTextual, or Tier4Unknown stale |
Previous commit summary, branch label, and public-safe age note. | Current-head behavior, current release status, or current proof path. | Reviewer or source owner. | This evidence belongs to an older source context and needs current-context confirmation. | Stop before current-head, current-release, or current-proof wording. | Validation checks |
| unknown evidence tier | unknown tier | Tier4Unknown |
Gap label, rule family, available artifact summary, and limitation. | Stronger evidence tier, semantic certainty, or complete coverage. | Reviewer or scanner owner. | The tier is unknown, so the statement must stay downgraded until evidence is classified. | Stop before tier upgrade or conclusion by confidence, repetition, or reviewer seniority. | Limitations and non-claims |
Safe conclusions
Say what the reduced evidence can support.
Review inputCoverage is reduced, so this is review input and the next proof question stays open.
Static referenceSyntax fallback found a static reference that needs semantic or owner review.
Missing artifactThe proof path is incomplete because public-safe artifact output is unavailable.
Private-only evidencePrivate-only evidence can support internal follow-up, but public wording waits for a public-safe summary.
Stale contextThe source context is stale, so current-head wording must wait.
Unknown tierThe tier is unknown, so the claim should be downgraded or held.
Unsafe conclusions
Reject patterns that upgrade a gap into a claim.
- Rejected pattern: reduced coverage proves there is no impact.
- Rejected pattern: a failed or reduced analysis is a clean repo.
- Rejected pattern: syntax fallback proves runtime behavior or compiler-resolved impact.
- Rejected pattern: the release is approved, certified, guaranteed, or safe because no public proof was found.
- Rejected pattern: operational safety, complete coverage, AI impact analysis, LLM analysis, prompt-based classification, embedding search, vector database analysis, autonomous approval, or replacement of human review is provided here.
Next evidence
Collect the public-safe proof target named by the row.
build/load failureCollect a build/load diagnostics summary or keep syntax fallback and analysis-gap labels attached.syntax fallbackCollect semantic-load results, owner review notes, or a public-safe syntax fallback report.missing semantic evidenceCollect unresolved-symbol context, project-load status, or a scanner limitation update.unsupported framework surfaceCollect a framework support note or rule catalog limitation before repeating framework wording.missing generated artifactPublish a generated public-safe report or choose a replacement proof path.private-only supportCreate a public-safe summary or keep the claim internal.stale commit contextCollect current source context before using current-head or release wording.unknown evidence tierReconcile the evidence tier or hold the conclusion at Tier4Unknown.Owner handoff
Transfer the evidence question, not approval.
Current labelState the row label, coverage label, and evidence tier exactly as the public-safe surface shows them.
Evidence availableName the public-safe evidence category and the missing evidence without quoting raw artifacts or private detail.
Requested next actionAsk the scanner owner, reviewer, build/tooling owner, framework owner, artifact publisher, site owner, or source owner to collect the next public-safe proof target.
Stop conditionKeep the claim downgraded, held, or internal until the proof target exists and the non-claim still travels with it.
Non-claimThe handoff does not approve a release, prove runtime behavior, assign blame, or replace human review.
Use the evidence gap register when the reduced label needs a row.
The evidence gap register records the follow-up fields for reduced coverage, missing proof paths, private-only support, stale source context, unsupported framework surfaces, validation gaps, and owner questions without upgrading the public claim.
Stop conditions
Stop before the wording outruns the evidence.
- Stop when the proof link is missing, unavailable, private-only, stale, or not public-safe.
- Stop when the statement drops the coverage label, evidence tier, rule ID or rule family, limitation, or owner follow-up.
- Stop when a reader asks for runtime behavior, production traffic, endpoint performance, outage cause, release approval, operational safety, or complete coverage.
- Stop when the next step requires source owner review, build/tooling repair, framework support work, artifact publication, or current source context.
Non-claims
This playbook does not publish raw proof or close the gap.
- No absence-of-impact proof, clean-repo claim under failed or reduced analysis, runtime proof, production traffic proof, endpoint performance proof, outage-cause proof, release approval, release safety, operational safety, or complete coverage proof.
- No AI impact analysis, LLM analysis, prompt-based classification, embedding search, vector database analysis, autonomous approval, or replacement of tests, source review, runtime observability, service owner review, or human review.
- No raw facts, raw SQLite content, analyzer logs, raw source snippets, raw SQL, config values, secrets, local paths, raw remotes, generated scan directories, private sample names, raw command output, hidden validation details, or credential-like values are public material.
- Artifact families such as fact streams, SQLite indexes, analyzer logs, scan manifests, reports, rule catalog entries, commit metadata, coverage labels, and limitations need public-safe summaries before public linking.
Adjacent surfaces
This page answers what to do after coverage is reduced.
Limitations and non-claimsCanonical boundary definitions; this playbook is the reduced-coverage action surface.
Validation checksCheck-result orientation; this playbook handles decisions after reduced coverage appears.
Static versus runtime boundariesRuntime question routing; this playbook stays with static evidence labels and handoff.
Stakeholder objection guideBroad objection answers; this playbook gives row-level wording, owner, and stop conditions.
Proof path FAQProof-path explanation; this playbook supplies the reduced-coverage scenario matrix.
Review claim checklistClaim repetition ritual; this playbook feeds checklist decisions when coverage is reduced.