Crate fallow_output

Expand description

Reusable output contract types for fallow.

This crate owns stable report DTOs that are not tied to CLI rendering. Higher-level output assemblers live above this crate in fallow-api or the CLI, while this crate remains the shared typed boundary those builders and non-CLI consumers can use. Shared selectors such as OutputFormat stay in fallow-types; this crate re-exports only the selectors needed to keep output builders on one contract surface.

Modules§

envelope: JSON-output envelope and utility types: SchemaVersion, ToolVersion, ElapsedMs, AuditIntroduced, plus the shared Meta, BaselineDeltas, BaselineMatch, RegressionResult, EntryPoints, and CheckSummary shapes referenced by every per-command envelope. The structs are always compiled (the JSON emission layer constructs them at runtime); the schemars::JsonSchema derive is gated per-struct on the schema feature. Typed envelope and utility-shape structs for the JSON output contract.
output: JSON-output augmentation types: IssueAction enum + variants. Schema-side counterpart of the augmentations the JSON layer adds to each dead-code finding. The structs are always compiled (typed dead-code wrappers in output_dead_code consume them at runtime); the schemars::JsonSchema derive is gated per-struct on the schema feature. Types that describe fallow’s JSON output contract.
output_dead_code: Typed envelope wrappers for the simple 1:1 dead-code findings (UnusedFile, PrivateTypeLeak, UnresolvedImport, CircularDependency, BoundaryViolation). Each wrapper flattens the bare finding via #[serde(flatten)] and carries a typed actions array populated at construction time, replacing the per-finding post-pass injection that previously grafted actions[] and introduced onto the schema. The introduced field is set by the audit pass via JSON map insertion and is None when serialized directly from Rust. The schemars::JsonSchema derive is gated per-struct on the schema feature. Typed envelope wrappers for the simple 1:1 dead-code findings whose actions are entirely determined by the wrapper type (no per-instance discriminants beyond what the bare finding already exposes).
output_health: Per-action types attached to health findings, hotspots, refactoring targets, and coverage-gap entries. Separated from the generic IssueAction tree in the output module so the health-specific variants live in a dedicated module. The structs are always compiled (the JSON emission layer constructs them through typed wrappers such as output_health::UntestedFileAction); the schemars::JsonSchema derive is gated per-struct on the schema feature. Per-action types attached to each health finding by the JSON output layer.

Structs§

AcceptedJudgment: One accepted judgment: the real anchored signal passed through with the agent’s framing FENCED as non-deterministic.
AffectedDirectory: One directory of the blast radius and how many affected files it holds.
AgentJudgment: One agent judgment.
AgentSchema: The shape the agent must return, embedded in the guide so a thin skill needs no frozen copy. Documents the anchoring + staleness contract in the wire.
AgentWalkthrough: The agent’s returned judgment JSON.
AuditNextStepsInput: Runtime-independent inputs for audit next steps.
AuditOutput: fallow audit --format json envelope.
AuditStylingContextLabelInput: Inputs for styling_audit_context_label, describing one styling rule’s gate state.
BoundariesListLogicalGroup: A pre-expansion autoDiscover logical group surfaced for observability. Captured during expansion so consumers can see the user-authored parent name and grouping intent after expansion would otherwise flatten it out of BoundariesListing::zones.
BoundariesListRule: A boundary import rule, expanded to operate on concrete child zone names after autoDiscover flattening. The user’s pre-expansion rule (keyed on the logical parent name, if any) is preserved on the corresponding BoundariesListLogicalGroup::authored_rule.
BoundariesListZone: A boundary zone after preset and autoDiscover expansion. Each entry classifies files into a single zone via glob patterns.
BoundariesListing: boundaries block carried by ListBoundariesOutput.
BranchingCognitive: The cognitive figure and what drove it.
BranchingFileDelta: One file’s contribution to the change.
BranchingMetric: One metric across the two revisions.
BranchingReport: The brief’s branching section.
BranchingScope: Size and composition of the compared set.
ChangeAnchor: One stable per-hunk CHANGE ANCHOR: a changed region the agent may cite as a judgment anchor IN ADDITION to a signal_id. Where a signal_id anchors a graph FINDING (“fallow emitted this exact finding”), a change_anchor anchors only a changed REGION (“fallow confirms this region changed”) , a strictly weaker guarantee, surfaced as anchor_kind on the accepted judgment so a consumer can tell the two apart. Graph/diff-derived; NEVER from prose.
CheckGroupedEntry: Single resolver bucket inside CheckGroupedOutput. Carries the group’s identifier, optional section owners, and a per-group flattened AnalysisResults.
CheckGroupedOutput: Envelope emitted by fallow dead-code --group-by ... --format json.
CheckOutput: Envelope emitted by fallow dead-code --format json (plus the check block inside the combined and audit envelopes).
CheckOutputInput: Inputs for building the dead-code JSON envelope.
CiIssue: Normalized CodeClimate issue used by CI comment renderers.
CiLocation: Source range attached to a normalized CI finding as supporting evidence.
ClockProvenance: The instant a run measured commit ages and staleness against.
CloneFamilyAction: Per-action wire shape attached to each CloneFamilyFinding. Mirrors the action types previously emitted by build_clone_family_actions: extract-shared, one apply-suggestion per RefactoringSuggestion on the family, and a trailing suppress-line.
CloneGroupAction: Per-action wire shape attached to each CloneGroupFinding and AttributedCloneGroupFinding (see crates/api/src/dupes_output.rs): extract-shared plus suppress-line. The typed wrappers replaced the legacy JSON post-pass injection that used to live in the CLI report layer.
CloneSiblingEvidence: A duplicate-code sibling referenced in target evidence.
CodeClimateIssue: Single CodeClimate-compatible issue inside CodeClimateOutput.
CodeClimateIssueInput: Fields needed to build one CodeClimate issue.
CodeClimateLines: Inclusive line range for CodeClimateLocation.
CodeClimateLocation: Location block inside CodeClimateIssue::location.
CodeClimateOutput: Envelope emitted by fallow --format codeclimate and fallow --format gitlab-codequality. GitLab Code Quality consumes the same shape. The wire form is a bare JSON array, not an object.
CombinedMeta: Optional _meta block for CombinedOutput.
CombinedNextStepsInput: Runtime-independent inputs for bare fallow combined next steps.
CombinedOutput: Bare fallow --format json envelope.
ComplexityViolation: Inner complexity-violation payload.
ComponentRollup: Component-level aggregate attached to a template complexity finding, pairing the template’s scores with the worst class-side function.
ContainmentEvent: A commit-gate containment event recorded by fallow impact.
ContributingFactor: A contributing factor that triggered or strengthened a recommendation.
ContributorEntry: One contributor row in ownership metrics.
CoordinationGapFact: One coordination-gap entry: a changed file exports symbols consumed by a consumer_file that is NOT in the diff. Deduped per (changed, consumer) pair (firing-precision rule R2).
CoverageAnalyzeOutput: Envelope emitted by fallow coverage analyze --format json.
CoverageGapSummary: Aggregate coverage-gap counters for the current analysis scope.
CoverageGaps: Static test coverage gaps derived from the module graph. Shows runtime files and exports with no test dependency path.
CoverageIntelligenceAction: Machine-actionable next step for a coverage-intelligence finding.
CoverageIntelligenceEvidence: Compact evidence values that led to a recommendation.
CoverageIntelligenceFinding: One combined coverage-intelligence finding.
CoverageIntelligenceReport: Combined coverage, runtime, complexity, and change-scope verdicts.
CoverageIntelligenceSummary: Aggregate metadata for coverage-intelligence output.
CoverageSetupFileToEdit: One manual edit the user must make to wire up coverage capture.
CoverageSetupMember: Per-workspace-member setup guidance inside CoverageSetupOutput::members.
CoverageSetupOutput: fallow coverage setup --json envelope.
CoverageSetupSnippet: Ready-to-paste code snippet accompanying a file edit.
CrossRepoImpactReport: The cross-repo aggregate report, fallow impact --all --format json.
CrossRepoProjectEntry: One project’s row in the cross-repo roll-up.
CrossRepoTotals: Grand totals across every tracked project (including repos whose directory no longer exists on disk: their past wins still count toward lifetime impact).
CssAnalyticsReport: Structural CSS analytics surfaced by fallow health --css.
CssAnalyticsSummary: Project-wide CSS analytics aggregates across every analyzed stylesheet (including stylesheets with no notable rule, which are not listed individually in files).
CssBlockOccurrence: One occurrence of a duplicate declaration block.
CssCandidateAction: A read-only verification step attached to a CSS cleanup candidate.
CssDuplicateBlock: A group of style rules across the project that share an identical declaration block: a copy-paste consolidation candidate (fallow’s duplication signal applied to CSS). Only blocks of 4+ declarations appearing in 2+ rules are reported, so the signal stays a strong copy-paste indicator rather than flagging legitimately-repeated small blocks.
CssFileAnalytics: Per-stylesheet CSS analytics.
CssNotationConsistency: A design-token notation-consistency candidate: the distinct notations used across the codebase for one value axis (today, length units on font-size), with a per-notation distinct-value count. Emitted only above a floor, since mixing notations for one axis is a “no single source of truth” smell. Advisory: the action is “standardize on one notation”, not a single search.
CssNotationCount: One notation bucket and the count of distinct values authored in it.
CvaDuplicateVariantBlock: A duplicated CVA / shadcn variant class block.
CvaVariantTokenDrift: A CVA / shadcn variant class value that can reuse an existing styling token.
CyclomaticPopulation: Disjoint populations feeding the cyclomatic distribution. Counts and sums across all three groups reconstruct its weighted mean. Module-scope units contribute to aggregates only and do not produce function findings.
CyclomaticUnitPopulation: Cyclomatic measurements for one kind of complexity unit.
DeadCodeNextStepsInput: Runtime-independent inputs for standalone dead-code next steps.
Decision: One consequential structural decision, framed as a judgment question for a human with taste, anchored to a fallow-emitted signal.
DecisionAction: A structured action attached to a surfaced decision (the agent-actionable surface). Mirrors the typed-action shape the rest of fallow emits.
DecisionSurface: The ranked, capped decision surface plus the set of signal_ids the deterministic layer emitted (the anti-hallucination allowlist).
DecisionSurfaceOutput: The separable decision-surface envelope: the single call that puts taste- decisions in front of a human, callable WITHOUT the full pipeline (the decision_surface MCP tool’s output). Carries kind/schema_version plus structured actions[] per decision.
DecisionSurfaceSchemaVersion: Independently-versioned wire-version newtype. Serializes as the integer DECISION_SURFACE_SCHEMA_VERSION.
DecisionWithActions: One decision plus its structured actions[].
DiffIndex: Parsed, command-neutral index of files and added lines in a unified diff.
DiffTriage: Stage 0 of the brief: triage facts derived purely from the diff size.
DirectCallerEvidence: A direct importer referenced in target evidence.
DirectCallerSymbolEvidence: Symbol details for a direct importer.
DirectionUnit: One directed review unit projected from the graph: a file the change touches, the concern to check, the out-of-diff consumers it must account for, and the routed expert. Graph-derived only (routing + impact closure), NEVER from prose.
DoctorCheck: One deterministic doctor check result.
DoctorOutput: Versioned readiness envelope emitted by fallow doctor --format json.
DoctorRemediation: Actionable remediation attached to a doctor check.
DoctorSummary: Counts for every per-check status.
DupesNextStepsInput: Runtime-independent inputs for standalone duplication next steps.
DupesOutput: Envelope emitted by fallow dupes --format json.
DupesOutputInput: Inputs for constructing a DupesOutput without exposing envelope assembly details to callers.
ErrorOutput: Structured JSON error emitted on stdout when --format json is active and a command fails. It carries no kind discriminator: it is distinguished from the kind-tagged success envelopes by the required error: true field, and is a document-root branch alongside FallowOutput and CodeClimateOutput in docs/output-schema.json. Agents that pass --format json and observe a non-zero exit code parse this shape from stdout.
EvidenceFunction: A function referenced in target evidence.
ExistingPrComment: Sticky comment already present on the PR, as found by a provider adapter.
ExplainOutput: Envelope emitted by fallow explain <issue-type> --format json.
FeatureFlagAction: Per-finding action emitted for feature flag findings.
FeatureFlagDeadCodeOverlap: Dead-code overlap block attached when a flag guards unused exports.
FeatureFlagFinding: One feature flag finding in JSON output.
FeatureFlagsConfidenceMeta: Feature flag confidence explanations.
FeatureFlagsKindMeta: Feature flag kind explanations.
FeatureFlagsMeta: Optional _meta block for FeatureFlagsOutput. Both fields are optional because the two contributors are independent: feature_flags details are present only with --explain, and telemetry is injected post-pass by attach_telemetry_meta whenever an analysis run id is available (which is the default path). Mirrors Meta / CombinedMeta, which also model telemetry as an optional, never-required property.
FeatureFlagsMetaDetails: Feature flag explanatory metadata.
FeatureFlagsOutput: Envelope emitted by fallow flags --format json.
FeatureFlagsOutputInput: Inputs for building fallow flags --format json.
FileHealthScore: Per-file health score combining complexity, coupling, and dead code metrics.
FixJsonOutput: JSON root emitted by fallow fix --format json.
FixJsonOutputInput: Inputs for building fallow fix --format json.
FocusMap: The weighted focus map: the ranked review-here units plus the FULL deprioritized escape-hatch list, so nothing is hidden.
FocusScore: The composite attention score, with the deterministic component sub-scores kept on the wire so the runtime layer adds its weight without recomputing the signals.
FocusUnit: One review unit on the focus map: its file, composite score, label, human reason, and any confidence flags.
FrameworkHealthDetector: Status for one framework-specific health detector.
FrameworkHealthDiagnostics: Framework-specific health detector coverage surfaced for agent consumers.
GateRunCounts: Recorded gate runs grouped by the gate that produced them. Local provenance only: the store never leaves the machine, so this answers “where do my gate runs come from”, never “how widely is fallow adopted”.
GitHubReviewComment: GitHub pull-request review comment.
GitLabReviewComment: GitLab merge-request discussion comment.
GitLabReviewPosition: position block inside GitLabReviewComment. Mirrors the GitLab merge-request discussion-position API.
GraphFacts: Stage 1 of the brief: graph-derived orientation facts.
GroupedReviewIssues: Review issues grouped per (path, line) for one-comment-per-location rendering.
HealthActionContext: Construction-time context for HealthFinding::with_actions.
HealthActionOptions: Options controlling how the action builder populates actions.
HealthActionsMeta: Auditable breadcrumb recording when health-finding suppress-line action hints were omitted from the report.
HealthBaselineStaleness: Staleness of a loaded health baseline.
HealthConfiguredThresholds: Threshold values configured by a single override entry.
HealthEffectiveThresholds: Resolved thresholds used to evaluate a health finding.
HealthFinding: Wire envelope for a single complexity finding.
HealthGroup: A health report scoped to a single group.
HealthGrouping: Wrapper carrying the resolver mode label alongside the partitioned groups.
HealthJsonOutputInput: Inputs for serializing a health report into the root JSON contract.
HealthNextStepsInput: Runtime-independent inputs for standalone health next steps.
HealthOutput: Envelope emitted by fallow health --format json (plus the health block inside the combined and audit envelopes).
HealthOutputInput: Inputs for constructing a HealthOutput without exposing envelope assembly details to callers.
HealthReport: Result of complexity analysis for reporting.
HealthScore: Overall project health score: 100 minus capped per-category penalties.
HealthScorePenalties: Per-component penalty breakdown for the health score.
HealthSummary: Summary statistics for the health report.
HealthTimings: Detailed timing breakdown for the health pipeline.
HealthTrend: Trend comparison between the current run and a previous snapshot. Shows per-metric deltas with directional indicators.
HotspotEntry: A hotspot: a file that is both complex and frequently changing.
HotspotFinding: Wire envelope for a single hotspot entry.
HotspotSummary: Scope metadata for the hotspot analysis.
ImpactClosureFacts: Stage 3 of the brief: the impact closure. The transitive affected-but-not-in-diff set plus the coordination gap. The differentiator a diff tool fundamentally cannot do, because it has no graph.
ImpactCounts: Per-category issue counts captured at a recorded run.
ImpactDigestCounts: Local impact digest counters used to render the impact-report next step.
ImpactReport: The rendered impact report, derived purely from the store.
InspectEvidence: evidence block of InspectOutput: one section per analysis.
InspectEvidenceSection: One evidence section: status, the scope the evidence covers, and either a data payload or a message explaining its absence.
InspectFileIdentity: Graph identity facts for a file target. Value-typed fields carry the graph’s verdict when analysis ran and are null when it did not.
InspectOutput: Envelope emitted by fallow inspect --format json.
InspectSymbolIdentity: Graph identity facts for a symbol target. Value-typed fields carry the graph’s verdict when analysis ran and are null when it did not.
IssueOutputContract: Output-facing contract metadata for a serialized dead-code result row.
LargeFunctionEntry: A function exceeding the very-high-risk size threshold (>60 LOC).
ListBoundariesOutput: Envelope emitted by fallow list --boundaries --format json. Surfaces the architecture boundary zones, rules, and the user’s pre-expansion autoDiscover logical groups so consumers can render grouping intent that expansion would otherwise flatten out of zones[].
ListEntryPointOutput: One entry point in fallow list --entry-points --format json.
ListOutput: Plain body emitted by fallow list --format json before an optional command-specific root envelope is attached.
ListPluginOutput: One active plugin in fallow list --plugins --format json.
NearDuplicateThemeToken: A Tailwind v4 @theme token that appears to duplicate an existing token by value. Emitted conservatively for comparable token namespaces, with the nearest existing token named so an agent has a concrete reuse target.
NearestStylingToken: A styling token candidate that can replace or explain a finding.
OwnershipMetrics: Ownership metrics for a hotspot file, derived from git history and CODEOWNERS declarations.
PartitionFacts: Stage 2 of the brief: the partition + order. The changed files split into coherent BY-MODULE units (the only byte-identical-deterministic clustering definition straight from the graph), plus a dependency-sensible review ORDER over those units (definitions before consumers, mechanical/leaf units last, ties broken by the path sort). Stage 2 sits UNDER the decision surface as a drill-down; it is the backbone the directed-review loop hands the agent.
PrCommentEnvelope: Rendered PR comment body plus posting signals for provider adapters.
PrCommentPostPlan: Posting decision handed to a provider adapter.
PrCommentPostPlanInput: Inputs for plan_pr_comment_post.
PrCommentRenderInput: Inputs for rendering a sticky PR/MR summary comment.
PrCommentTruncation: Finding-count truncation metadata for a rendered PR comment.
PrDecisionAnnotation: File-anchored annotation inside PrDecisionSurface::annotations.
PrDecisionDetails: Report content and links inside PrDecisionSurface::details.
PrDecisionGate: One quality-gate outcome inside PrDecisionSurface::gates.
PrDecisionSurface: Provider-neutral PR gate decision, consumed by CI integrations to publish a check run or equivalent status surface.
PrDetailsArtifact: Full-findings report artifact backing the PR summary comment’s details link, grouped into per-area sections.
PrDetailsRow: One finding row inside PrDetailsSection::rows.
PrDetailsSection: One findings section inside PrDetailsArtifact::sections.
PrStatusContext: One commit-status context derived from a PrDecisionSurface.
PrSummaryArea: One analysis area row in the PR summary table.
PrSummaryFinding: One finding row in the PR summary’s top-findings list.
PrSummaryInput: Inputs for render_pr_summary.
RawStyleValue: A located raw CSS declaration value on a scale-sensitive styling axis.
ReactHookProfile: Per-component React hook profile derived from the cached hook_uses IR at the health layer. Descriptive context that refines the bare ComplexityViolation::react_hook_count headline with a per-kind breakdown and the maximum useEffect dependency-array arity.
RefactoringTarget: One prioritized refactoring recommendation in the health report’s refactoring_targets section.
RefactoringTargetFinding: Wire envelope for a single refactoring target.
RejectedJudgment: One rejected judgment plus the reason it was rejected.
RenderFanInTopComponent: One located high-fan-in React/Preact component for the descriptive top_render_fan_in blast-radius list on VitalSigns.
ResolutionEvent: A resolved or suppressed finding attribution event.
ReviewBriefHeader: Informational audit metadata carried by the review brief wire envelope.
ReviewBriefOutput: The full fallow audit --brief --format json envelope. Carries the informational verdict, the triage and graph-facts orientation stages, plus the reused “subtract” section (the same dead-code / duplication / complexity payload fallow audit --format json emits).
ReviewBriefSchemaVersion: Independently-versioned wire-version newtype for the brief envelope. Serializes as the integer REVIEW_BRIEF_SCHEMA_VERSION.
ReviewBriefSubtractSections: CLI-built audit subreports that are embedded in the audit brief envelope.
ReviewBriefWireOutput: Complete fallow audit --brief --format json wire envelope.
ReviewCommentRenderInput: Inputs for rendering one inline review comment from a location group.
ReviewDeltas: Diff-aware deterministic deltas (6.A), framed new-vs-pre-existing against the audit base snapshot. Each entry is a brief summary/verdict line.
ReviewDirection: The review direction artifact: the order to review in, the coherent units, and per-unit concern lens + out-of-diff + expert. A minimal projection of the EXISTING graph facts (routing units + impact closure); the full weighted-focus engine is a later epic. Graph-derived only (injection-resistant).
ReviewEnvelopeMeta: meta block inside ReviewEnvelopeOutput.
ReviewEnvelopeOutput: Envelope emitted by fallow --format review-github / review-gitlab.
ReviewEnvelopeRenderInput: Inputs for rendering a GitHub/GitLab review envelope.
ReviewEnvelopeRenderResult: Rendered review envelope plus side-channel signals for CLI telemetry.
ReviewEnvelopeSummary: Summary block on ReviewEnvelopeOutput.
ReviewEnvelopeTruncation: Truncation signals produced while rendering a review envelope.
ReviewGitlabDiffRefs: GitLab diff refs for a review-envelope position.
ReviewId: Stable identifier used to isolate independent review integrations on the same pull or merge request.
ReviewReconcileOutput: Envelope emitted by fallow ci reconcile-review --format json. Used by CI integrations to drive comment carry-over and stale-comment cleanup across PR / MR revisions.
ReviewUnitFact: One review unit: a coherent by-module cluster of the changed set.
RiskProfile: Risk profile: percentage of functions in each risk bin.
RoutingFacts: The full routing section: one unit per changed source file with a routable signal. Files with no ownership signal are omitted (no noise).
RoutingUnit: One routed unit with its experts and bus-factor flag.
RuntimeCoverageAction: Suggested follow-up action for a runtime coverage finding.
RuntimeCoverageBlastRadiusEntry: One blast-radius entry in runtime_coverage.blast_radius: how far a change to the function would ripple.
RuntimeCoverageCaptureQuality: Quality-of-capture signals emitted by the sidecar so the CLI can explain short-window captures honestly instead of letting users blame the tool.
RuntimeCoverageDiscriminators: Discriminator inputs that PRODUCED a finding’s verdict (fallow-rs/fallow-cloud#321), emitted alongside the verdict so an agent can reproduce it and see the minimum-observation confidence cap instead of re-deriving them from scratch. F4: these make the EXISTING Fallow-owned discriminators legible; they are not a new or external signal and gate nothing. Pairs with evidence.static_status (the static half of the discriminator set).
RuntimeCoverageEvidence: Supporting evidence for a finding (mirrors fallow_cov_protocol::Evidence).
RuntimeCoverageFinding: One per-function runtime-coverage finding in runtime_coverage.findings.
RuntimeCoverageHotPath: One hot function in runtime_coverage.hot_paths, ranked by invocations.
RuntimeCoverageImportanceEntry: One production-importance entry in runtime_coverage.importance, scoring how much a function matters in production.
RuntimeCoverageMessage: Non-fatal diagnostic emitted while merging runtime coverage.
RuntimeCoverageProvenance: Provenance of a runtime-coverage report (fallow-rs/fallow-cloud#319), mirroring the cloud runtime-context provenance block so the local-capture and cloud surfaces present one portable shape. F4: provenance is context only; it never gates a verdict or confidence.
RuntimeCoverageReport: Runtime coverage findings merged into the health report or emitted by fallow coverage analyze. Present in health output when –runtime-coverage is used. Shape mirrors the runtime coverage JSON contract; cloud mode fetches runtime facts explicitly and merges them locally with AST/static analysis.
RuntimeCoverageSummary: Summary block mirroring fallow_cov_protocol::Summary (0.3 shape).
SarifDocumentInput: Fields needed to build a SARIF document envelope.
SarifFindingFields: Intermediate fields extracted from one issue for SARIF result construction.
SarifFindingInput: Normalized finding input for output-owned SARIF result assembly.
SarifResultInput: Fields needed to build one SARIF result object.
SarifRuleInput: Fields needed to build one SARIF rule object.
SarifSourceSnippetCache: Lazily reads source files so SARIF result builders can attach stable line snippets.
ScopedUnusedClasses: A Vue SFC’s <style scoped> classes that appear nowhere else in the component (cleanup candidates).
SecurityBlindSpotFile: One file inside a blind-spot group.
SecurityBlindSpotGroup: One actionable blind-spot group.
SecurityBlindSpotsOutput: The fallow security blind-spots --format json envelope.
SecurityBlindSpotsSummary: Aggregate counts for blind-spot output.
SecurityGate: The gate block on SecurityOutput, present only when --gate <mode> ran. Invariant: verdict == Fail IFF exit code 8 IFF new_count > 0.
SecurityOutput: The fallow security --format json envelope. FallowOutput discriminates it by the kind: "security" tag; the optional gate block is additive and is not part of that discrimination.
SecurityOutputConfig: Allowlisted config context for fallow security --format json.
SecurityOutputRulesConfig: Per-rule severity context inside SecurityOutputConfig::rules.
SecurityReachabilityCounts: Fixed reachability counters for summary JSON.
SecurityRuleMeta: Output-facing metadata for one security rule.
SecurityRuleSeverityConfig: Configured-versus-effective severity for one security rule.
SecurityRuntimeStateCounts: Fixed runtime coverage counters for summary JSON.
SecuritySeverityCounts: Fixed severity counters for summary JSON.
SecuritySummary: Aggregate counts for fallow security --summary --format json.
SecuritySummaryOutput: Compact fallow security --summary --format json payload. Uses the same kind: "security" discriminator as the full payload, but omits candidate arrays and exposes only aggregate counts.
SecuritySurvivor: One verifier-retained candidate row.
SecuritySurvivorsOutput: The fallow security survivors --format json envelope.
SecuritySurvivorsSummary: Aggregate counts for survivor rendering.
SecurityUnresolvedCalleeDiagnostics: Bounded unresolved-callee diagnostics for fallow security --format json.
SecurityUnresolvedCalleeReasonCount: Count of unresolved callees for one reason.
SecurityUnresolvedCalleeSample: One sampled unresolved-callee row.
SecurityUnresolvedCalleeTopFile: Count of unresolved callees in one file.
SecurityVerifierVerdict: One supported verifier verdict input row.
SimilarCodeAction: Read-only follow-up exposed for a candidate.
SimilarCodeCacheClearOutput: Result of explicitly clearing the derived project-namespaced vector cache.
SimilarCodeCacheSummary: Privacy-safe cache accounting. Source fragments are never represented.
SimilarCodeCandidate: One unverified semantic similar-code candidate.
SimilarCodeCandidateSnapshot: Bounded handoff for inspecting one immutable discovery candidate without rerunning global retrieval or ranking.
SimilarCodeCompletion: Typed completion, limit, skip, and cache accounting.
SimilarCodeDiagnostic: Actionable diagnostic without a severity or gate implication.
SimilarCodeEnrichmentAvailability: Availability of every supported source-grounded enrichment.
SimilarCodeGeneration: Complete provenance needed to reproduce candidate generation.
SimilarCodeGenerationParameters: Parameters that materially affect generated embeddings and scores.
SimilarCodeInspectOutput: fallow similar-code inspect --format json output.
SimilarCodeInspectPacket: Bounded source-grounded packet for one immutable candidate.
SimilarCodeLimits: Effective resource limits for a similar-code run.
SimilarCodeLocation: Exact named location of one candidate function.
SimilarCodeModelProvenance: Immutable model artifact provenance.
SimilarCodeNamedReference: One named graph reference used in an inspect packet.
SimilarCodeOutput: Raw fallow similar-code --format json output.
SimilarCodePhaseCompletion: Accounting for one bounded generation phase.
SimilarCodeProviderProvenance: Immutable local provider provenance for one generation run.
SimilarCodeReviewOutput: fallow similar-code review --format json output.
SimilarCodeReviewProvenance: Digests that make the review join reproducible without exposing source.
SimilarCodeReviewedCandidate: One candidate joined with its separate verdict, if safely matched.
SimilarCodeScopeProvenance: Effective endpoint scope used for corpus admission and pair retention.
SimilarCodeSideEvidence: Bounded evidence for one side of an inspect packet.
SimilarCodeSkip: Count of skipped work for a stable reason.
SimilarCodeStatusOutput: Machine-readable readiness of the exact local companion and pinned model.
SimilarCodeVerdict: Separate verdict input for one immutable candidate.
SimilarCodeVerdictInput: Versioned external verdict document consumed by review.
SplitInPlace: One file present on both revisions whose branching held while it gained functions and its largest function shrank.
StylingFinding: One advisory STYLING FINDING: the graduation of a descriptive css candidate into a first-class, severity-aware, suppressible finding surfaced in fallow audit. The styling domain’s OWN finding type (not borrowed into the dead-code AnalysisResults, and not glued in the CLI). code is the kebab IssueKind code (e.g. css-token-drift), so severity / inline suppression / SARIF / MCP all resolve via the shared issue_meta contract through IssueKind::parse(code). One Vec<StylingFinding> carries every styling family; the code discriminates.
StylingHealth: Project-level styling-health score: a SECOND health axis computed purely from the structural CSS analytics (CssAnalyticsReport), orthogonal to the JS/TS code-health HealthScore. Surfaced only alongside the --css analytics, so a plain fallow health run is byte-unchanged. The code score and grade stay untouched: styling health is additive, never folded into the code score.
StylingHealthPenalties: Per-category penalty breakdown for the styling-health score. Each field is the number of points subtracted from a starting 100 for one CSS signal family, already capped at its category ceiling. A 0.0 field means “the signal was evaluated and clean”; the whole struct is only ever built when CSS analytics were produced, so there is no “missing pipeline” ambiguity to model with Option here (the parent StylingHealth is itself Option on the report).
SuppressionInventoryEntry: One suppression marker in the inventory.
SuppressionInventoryFile: One file’s suppression listing.
SuppressionInventoryOutput: The fallow suppressions --format json envelope. FallowOutput discriminates it by the kind: "suppression-inventory" tag.
SuppressionInventoryOutputInput: Inputs for building fallow suppressions --format json.
SuppressionInventorySummary: Project-level totals for the suppression inventory.
SuppressionKindCount: One by_kind row in the suppression inventory summary.
TailwindArbitraryValue: A distinct Tailwind arbitrary-value utility token used in markup, with its total use count and first location (a design-token-bypass candidate).
TargetEvidence: Evidence linking a target back to specific analysis data.
TargetThresholds: Adaptive thresholds used for refactoring target scoring.
ThresholdOverrideMetrics: Current complexity metrics for a matched threshold override entry.
ThresholdOverrideState: Report entry describing whether a threshold override is active, stale, or no longer matching any analyzed file or function.
TokenConsumerLocation: Where one Tailwind or CSS-in-JS design token is consumed, and through which surface. One entry in a TokenConsumers::consumers sample.
TokenConsumers: A location-aware reverse index of where one design token is consumed, so an agent editing the token can see its blast radius before changing or removing it. Covers TWO token origins. The always-available discriminator is the token SHAPE: a Tailwind token is the ---prefixed custom property (--color-brand), a CSS-in-JS token is a dotted access path with no -- prefix (vars.color.primary). The per-consumer kind also discriminates origin, but only when consumer_count > 0 (a consumer_count: 0 entry has an empty consumers array and thus no kind), so branch on the token prefix for the zero-consumer case. The two origins:
TraceUnusedExportInput: Deterministic unused-export trace target selected by the caller.
TrendCount: Raw numerator/denominator for a percentage metric.
TrendMetric: A single metric’s trend between two snapshots.
TrendPoint: A reference to a snapshot used in trend comparison.
TrendSummary: A computed trend between the two most recent records.
TruncationNote: A note for decisions collapsed below the cap.
TsAliasMeta: TypeScript backwards-compat alias emitted for a dead-code result row.
TypeAwareStatusOutput: Envelope emitted by fallow type-aware status --format json.
UndefinedKeyframes: An animation reference (animation / animation-name) to a @keyframes name that is defined in no stylesheet anywhere in the project (the “used-but-undefined” direction). Usually a typo or a removed animation; occasionally a @keyframes defined in CSS-in-JS the CSS parser never sees.
UnreferencedCssClass: A global CSS class defined in a plain .css/.scss rule whose literal name is referenced by no in-project markup (the CSS analogue of an unused export). A heavily-gated candidate, never a gated finding: the class may be applied from an HTML email, server template, CMS, or Markdown the parser never sees.
UnreferencedKeyframes: A @keyframes defined in a stylesheet but referenced by no animation in any stylesheet (cleanup candidate).
UnresolvedClassReference: A static class / className token in markup that matches no CSS class defined anywhere in the project but is one edit away from a class that IS defined (a likely typo or stale rename). The CSS analogue of an unresolved import. A candidate, never a gated finding: the token could be defined in CSS-in-JS or an external stylesheet the parser never sees.
UntestedExport: Runtime export that no test-reachable module references.
UntestedExportFinding: Wire-shape envelope for an UntestedExport finding. Same pattern as UntestedFileFinding: flattens the bare finding and carries a typed actions array computed at construction time.
UntestedFile: Runtime code that no test dependency path reaches.
UntestedFileFinding: Wire-shape envelope for an UntestedFile finding. Carries the bare UntestedFile flattened in plus a typed actions array. The action vec is computed at construction time using a project-root-relative path so descriptions match strip_root_prefix’s post-pass output on the inner path field. Schemars derives the merged shape natively; this retires the augment_finding_definition graft for UntestedFile.
UnusedAtRule: An unused CSS at-rule entity (an @property registration with no var() reference, or an @layer declaration never populated), located by its first definition. A cleanup candidate, never a gated finding.
UnusedFontFace: An @font-face family declared in a stylesheet but referenced by no font-family anywhere in the project: a dead web-font payload. A cleanup candidate (the family could be applied from inline styles or JavaScript).
UnusedThemeToken: A Tailwind v4 @theme design token defined in a stylesheet whose generated utility, var() reads, and arbitrary-value references appear nowhere in the project: a dead design token (the unused-export of the token era). A candidate, never a gated finding: the token could be consumed by a Tailwind plugin, a published design-system surface, or a non-CSS-aware build step the scan cannot see (those cases are gated out before this is emitted).
VitalSigns: Project-wide vital signs: a fixed set of metrics for trend tracking.
VitalSignsCounts: Raw counts backing the vital signs percentages.
VitalSignsSnapshot: A point-in-time snapshot of project vital signs, persisted to disk.
WalkthroughAccounting: Honest file accounting for the walkthrough header + status, reconciled so the numbers add up: staged + cleared + excluded == changed.
WalkthroughGuide: The fallow review --walkthrough-guide envelope: the current digest + schema the agent fetches. The tool owns this; the skill stays thin (it fetches this rather than embedding a frozen copy). Always emitted with exit 0.
WalkthroughValidation: The fallow review --walkthrough-file validation envelope: the result of post-validating the agent’s judgment against the live graph. Always exit 0.
WeakeningSignal: One weakening signal: a category, the file it was detected in, and a short human-readable evidence string. Reviewer-private; never gates.
WorkspaceInfo: One workspace package emitted by fallow workspaces --format json.
WorkspacesOutput: fallow workspaces --format json envelope.

Enums§

AuditCommand: Audit command singleton carried by AuditOutput.
AuditDisplayGate: Which findings an audit rule gates on.
AuditDisplaySeverity: Effective severity of an audit rule as shown in rendered output.
CiProvider: Supported CI review providers for generated comments.
ClockSource: Where the run’s reference epoch came from.
CloneFamilyActionType: Discriminant for CloneFamilyAction::kind.
CloneGroupActionType: Discriminant for CloneGroupAction::kind. Mirrors the action types emitted by the legacy build_clone_group_actions walker.
CodeClimateAnnotationField: Optional grouped CodeClimate annotation field.
CodeClimateIssueKind: Discriminator value for CodeClimateIssue::kind.
CodeClimateSeverity: CodeClimate severity scale.
CognitiveAttribution: Where a cognitive-complexity improvement came from.
Confidence: Confidence level for a refactoring recommendation.
ConfidenceFlag: A per-unit confidence flag. The EXACT panel-decided strings: a dynamically- wired or re-export-heavy unit carries one so its static-reachability signal is not trusted as complete (the anti-silent-de-prioritization guard). The flag NEVER lowers the score; it is advisory provenance.
ConsumerKind: The surface through which a design token is consumed. The theme-var / css-var / utility / apply kinds are Tailwind v4 @theme consumption; the js-member / js-call kinds are CSS-in-JS consumption (member access through a same-file or imported StyleX/vanilla-extract token binding, a StyleX theme-group call, or a PandaCSS token-path call). The kind is the disjoint origin signal that distinguishes a Tailwind token entry from a CSS-in-JS token entry in the shared token_consumers list.
ContributorIdentifierFormat: Encoding applied to a ContributorEntry::identifier.
CoverageAnalyzeSchemaVersion: Schema-version discriminator for CoverageAnalyzeOutput.
CoverageIntelligenceConfidence: Confidence in the joined evidence and resulting recommendation.
CoverageIntelligenceMatchConfidence: Confidence tier for the cross-surface evidence match.
CoverageIntelligenceRecommendation: Recommended action family for a combined finding.
CoverageIntelligenceSchemaVersion: Coverage-intelligence JSON contract version. Scoped to the coverage_intelligence block and independent of the top-level fallow JSON schema_version.
CoverageIntelligenceSignal: Ordered evidence signals behind a coverage-intelligence finding.
CoverageIntelligenceVerdict: Headline verdict for the combined coverage-intelligence report.
CoverageModel: Coverage model used for CRAP score computation.
CoverageSetupFramework: Framework detected during coverage setup; drives which instrumentation guidance is emitted.
CoverageSetupPackageManager: Package manager detected from the project’s lockfile.
CoverageSetupRuntimeTarget: Runtime environment coverage capture must instrument.
CoverageSetupSchemaVersion: Schema-version discriminator for CoverageSetupOutput.
CoverageSource: Provenance of a CRAP finding’s coverage signal.
CoverageSourceConsistency: Whether CRAP findings in the report used one coverage-source kind or a mix.
CoverageTier: Coverage tier classification for CRAP findings.
CrossRepoImpactSchemaVersion: Independent wire-version for the cross-repo report, on its own cadence (it versions separately from the per-project ImpactReportSchemaVersion and the on-disk STORE_SCHEMA_VERSION).
CssCandidateActionType: Discriminant for CssCandidateAction::kind.
DecisionActionType: The discriminated action kinds a decision can carry.
DecisionCategory: The exactly-three shippable decision categories (the SOLID-3). No cut category (abstraction / deletion / convention / irreversibility) is representable: this enum is the structural guarantee that confirmed-noise categories never ship.
DoctorCheckCategory: Stable category for a doctor check.
DoctorCheckId: Stable identifier for a doctor check. Declaration order is output order.
DoctorCheckStatus: Per-check readiness outcome.
DoctorStatus: Aggregate readiness outcome.
EffortEstimate: A ranked refactoring recommendation for a file.
EnabledSource: Why Impact tracking is (or is not) active for a project. Project = an explicit per-repo enable; User = the user-global default with no per-repo decision; Default = off (no per-repo decision and no global default).
ExceededThreshold: Which complexity threshold was exceeded.
FallowOutput: Typed root of every fallow JSON envelope shape that serializes as a JSON object and participates in the documented FallowOutput contract. The schema derived from this enum drives the document-root oneOf in docs/output-schema.json.
FeatureFlagActionType: Feature flag action discriminants.
FeatureFlagConfidence: Feature flag confidence values emitted in JSON.
FeatureFlagKind: Feature flag kind values emitted in JSON.
FindingSeverity: Severity tier indicating how far a function exceeds complexity thresholds.
FocusLabel: The focus label for a review unit. Skip is the SAFE explicit-skip label and is runtime-backed ONLY: it is producible solely on the paid runtime path and solely for a unit runtime-proves cold with zero risk signals. Free mode (no runtime evidence) emits only ReviewHere / NotPrioritized, never Skip (the build-focus-map labeller cannot reach the Skip arm without runtime input), so the free-tier “rank but never skip” stance holds by construction.
FrameworkHealthDetectorStatus: Detector status codes for framework health observability.
GitHubReviewSide: Singleton side discriminator for GitHubReviewComment::side.
GitLabReviewPositionType: Singleton position-type discriminator for GitLabReviewPosition.
GroupByMode: Resolver mode label for grouped envelopes (dead-code, dupes, health).
ImpactReportSchemaVersion: Wire-version discriminator for ImpactReport. Independent from the global SchemaVersion (the impact report versions on its own cadence) and from the on-disk STORE_SCHEMA_VERSION (the persisted store shape versions separately). Serializes as a string const so JSON consumers can switch on it, matching the other independently-versioned envelopes (e.g. CoverageAnalyzeSchemaVersion).
ImpactTrendDirection: Direction of a count trend between two recorded runs.
InspectEvidenceScope: Granularity an InspectEvidenceSection payload covers.
InspectIdentity: identity block of InspectOutput; shape follows the target type.
InspectSectionStatus: Status of an InspectEvidenceSection.
InspectTargetDescriptor: target block of InspectOutput, tagged by type.
OutputFormat: Output format for fallow results.
OwnershipState: Ownership lifecycle state of a file.
PrCommentLayout: Body layout variant for the sticky PR comment.
PrCommentPostAction: What the provider adapter should do with the rendered comment.
PrCommentPostSkipReason: Why a post was skipped.
PrDecisionAnnotationLevel: Display level for a PrDecisionAnnotation; values mirror GitHub check-run annotation levels.
PrDecisionConclusion: Outcome of the overall decision or a single gate; values mirror GitHub check-run conclusions.
PrStatusMode: How a decision surface fans out into commit-status contexts.
PrSummaryScope: What slice of the codebase the summarized run analysed.
PrSummaryStatus: Per-area gate status shown in the PR summary table.
RecommendationCategory: Category of refactoring recommendation.
ReviewCheckConclusion: meta.check_conclusion for the GitHub review envelope. Maps to the GitHub Checks API conclusion field.
ReviewComment: Per-line review comment. Schema is an anyOf between GitHub and GitLab shapes; at runtime every entry in a single envelope comes from the same provider because the envelope is built from one provider’s branch in crates/cli/src/report/ci/review.rs::render_review_envelope.
ReviewEffort: Suggested reviewer effort, a pure function of RiskClass.
ReviewEnvelopeEvent: Singleton GitHub review-event marker.
ReviewEnvelopeSchema: Schema-version discriminator for the review envelope.
ReviewProvider: Review-envelope provider tag.
ReviewReconcileSchema: Schema-version discriminator for the review reconcile envelope.
RiskClass: Coarse risk classification for a changeset, a pure function of the change size (file count plus, once threaded, net lines).
RootEnvelopeMode: JSON root envelope discriminator policy.
RuntimeCoverageConfidence: Confidence level for a runtime coverage finding.
RuntimeCoverageDataSource: Runtime coverage source used to produce the summary.
RuntimeCoverageReportVerdict: Top-level verdict for the whole runtime-coverage report. Mirrors fallow_cov_protocol::ReportVerdict. The verdict is the SINGLE most actionable finding; for the full set of findings see RuntimeCoverageReport::signals. The verdict promotes hot-path-touched above cold-code-detected in PR-review context (when the CLI was given a change-scope: --diff-file or --changed-since) because the touched-hot-path is event-tied to the current diff and reviewers need it to be the top-line signal. In standalone analysis (no change scope), cold-code-detected remains primary.
RuntimeCoverageRiskBand: Blast-radius risk band. The current thresholds are high at >=20 static callers or >=1,000,000 traffic-weighted caller reach; medium at >=5 callers or >=50,000 weighted reach; low otherwise.
RuntimeCoverageSchemaVersion: Runtime coverage JSON contract version. This is scoped to the runtime_coverage block and is independent of the top-level fallow JSON schema_version.
RuntimeCoverageSignal: Discrete signal captured during runtime-coverage post-processing. verdict collapses to one summary value; signals enumerates ALL findings the report carries so JSON consumers, CI dashboards, and agents can reason about them independently of the headline. Order is stable: severity-descending so the first entry mirrors a sensible non-PR-context verdict.
RuntimeCoverageVerdict: Protocol-level per-function runtime coverage verdict derived from the decision table in fallow-cov-protocol. The CLI’s runtime_coverage.findings array omits active entries even though the underlying enum still includes it.
RuntimeCoverageWatermark: License or trial watermark applied to runtime coverage output.
SecurityBlindSpotsSchemaVersion: The fallow security blind-spots --format json schema version.
SecurityGateVerdict: Gate verdict on the wire. fail is the CI-state token; human output renders it as “REVIEW REQUIRED” because these stay unverified candidates, never confirmed vulnerabilities.
SecuritySchemaVersion: The fallow security --format json schema version. Independently versioned from the main contract, mirroring ImpactReportSchemaVersion.
SecuritySurvivorsSchemaVersion: The fallow security survivors --format json schema version.
SecurityVerifierVerdictStatus: Verifier verdict status accepted by fallow security survivors.
SimilarCodeActionType: Read-only actions supported by the candidate workflow.
SimilarCodeCacheClearSchemaVersion: Version singleton for vector-cache clear output.
SimilarCodeCacheStatus: Vector cache outcome for a run.
SimilarCodeCompletionStatus: Overall trustworthiness of an emitted result set.
SimilarCodeDiagnosticDomain: Non-severity diagnostic domain for similar-code generation.
SimilarCodeDomainOutcome: Domain outcome assigned by review, never by candidate generation.
SimilarCodeEnrichmentState: Explicit availability state for one optional enrichment source.
SimilarCodeInspectSchemaVersion: Version singleton for a similar-code inspect packet.
SimilarCodePhase: Bounded phase names in the similar-code pipeline.
SimilarCodePhaseStatus: Completion state for one generation phase.
SimilarCodeProvider: Provider families admitted by the version 1 public contract.
SimilarCodeReviewSchemaVersion: Version singleton for reviewed similar-code output.
SimilarCodeSchemaVersion: Version singleton for raw similar-code output.
SimilarCodeSideEffectHint: Conservative syntactic side-effect hint for an inspected function.
SimilarCodeSimilarityBand: Stable, coarse interpretation of a candidate score.
SimilarCodeSkipReason: Stable reasons why admitted work was skipped or truncated.
SimilarCodeStatusSchemaVersion: Version singleton for local-provider status output.
SimilarCodeVerdictMatch: How review matched an external verdict to the current candidate.
SimilarCodeVerdictSchemaVersion: Version singleton for the separate human or agent verdict document.
SimilarCodeVerdictValidationError: Validation failures for the independent verdict judgments.
SimilarCodeVerificationStatus: Verification state of a raw semantic candidate.
StylingAgentDisposition: Agent handling hint for a StylingFinding.
StylingFindingConfidence: Confidence hint for a StylingFinding.
StylingFindingSeverity: Effective configured severity for a styling finding.
StylingHealthConfidence: Trust level for a StylingHealth grade. TWO variants (not the three-tier high/medium/low of crate::Confidence / FeatureFlagConfidence) ON PURPOSE: styling confidence is binary (the grade is either reliable for the analyzed surface or it is not), not three distinct evidence tiers, so a never-emitted Medium would be dead surface. Serializes lowercase ("high" / "low"), matching the sibling confidence enums’ vocabulary.
SuppressionInventoryLevel: Suppression marker scope.
SuppressionInventoryOrigin: How a suppression in the inventory was authored.
SuppressionInventorySchemaVersion: The fallow suppressions --format json schema version. Independently versioned from the main contract, mirroring SecuritySchemaVersion.
TestAdjacency: Whether a changed source unit has a test file importing it, and whether that test moved with the change. A direct-importer fact from the graph, not a coverage claim: untouched says a test exists and was not edited, which is the verification question the reviewer asks the author, never an answer.
ThresholdOverrideDimension: Which threshold dimension a thresholdOverrides state row describes.
ThresholdOverrideStatus: Lifecycle state for a configured threshold override.
ThresholdSource: Source for a finding’s effective thresholds.
TrendDirection: Direction of a metric’s change, semantically (improving/declining/stable).
UnusedAtRuleKind: Discriminant for UnusedAtRule::kind.
WeakeningKind: The category of a single weakening signal.

Constants§

ACTIONS_AUTO_FIXABLE_FIELD_DEFINITION: _meta description for the per-action auto_fixable bool.
ACTIONS_FIELD_DEFINITION: _meta description for the per-finding actions[] array shared across JSON output.
AFFECTED_DIR_CAP: Maximum number of directories reported in ImpactClosureFacts::affected_by_dir. Directories beyond the cap are the lightest ones and are counted in ImpactClosureFacts::affected_by_dir_omitted.
AFFECTED_SAMPLE_CAP: Maximum number of affected-but-not-in-diff paths sampled into ImpactClosureFacts::affected_not_shown.
AUDIT_SCHEMA_VERSION: Current schema version for fallow audit --format json.
CHECK_DOCS: Docs URL for the dead-code/check command.
CHECK_SCHEMA_VERSION: Current schema version for the dead-code/check JSON envelope.
COGNITIVE_EXTRACTION_THRESHOLD: Cognitive complexity at or above which a function is flagged as an extraction candidate in refactor targets and cited as a contributing factor on file health scores.
COMBINED_SCHEMA_VERSION: Current schema version for bare combined JSON output.
COVERAGE_ANALYZE_DOCS: Docs URL for fallow coverage analyze --format json --explain.
COVERAGE_SETUP_DOCS: Docs URL for the runtime coverage setup command’s agent-readable JSON.
DECISION_SURFACE_SCHEMA_VERSION: Wire version for the fallow decision-surface --format json envelope.
DEFAULT_BRANCHING_TOLERANCE: How many branch points a total may move before the change is treated as flat. Absolute rather than proportional: a small changeset is where a two-branch move matters most, and a proportional band would blind exactly that case.
DEFAULT_COGNITIVE_CRITICAL: Default cognitive complexity threshold for “critical” finding severity.
DEFAULT_COGNITIVE_HIGH: Default cognitive complexity threshold for “high” finding severity; crossing it upgrades a complexity finding from moderate to high.
DEFAULT_CRAP_CRITICAL: CRAP score threshold for “critical” severity. CC=10 untested gives 110, CC=12 untested gives 156; 100 lands between the two and flags genuinely dangerous combinations of high complexity and low coverage.
DEFAULT_CRAP_HIGH: CRAP score threshold for “high” severity. CC=7 untested -> 56, CC=10 -> 110.
DEFAULT_CYCLOMATIC_CRITICAL: Default cyclomatic complexity threshold for “critical” finding severity. Also the cutoff for the v2 health score’s critical-complexity density penalty (share of functions at or above this value).
DEFAULT_CYCLOMATIC_HIGH: Default cyclomatic complexity threshold for “high” finding severity; crossing it upgrades a complexity finding from moderate to high.
DEFAULT_MAX_UNIT_SIZE: Default unit-size ceiling (health.maxUnitSize): functions over 60 lines of code are reported as oversized. Mirrors the config crate’s default so renderers can fill an effective-thresholds fallback without a config handle.
DOCTOR_SCHEMA_VERSION: Current schema version for fallow doctor --format json.
DUPES_DOCS: Docs URL for the duplication command.
DUPES_PROGRAMMATIC_SCHEMA_VERSION: Current schema version for programmatic duplication JSON.
DUPES_SCHEMA_VERSION: Current schema version for fallow dupes --format json.
FEATURE_FLAGS_SCHEMA_VERSION: Current schema version for feature-flag JSON output.
GHAS_SARIF_FINGERPRINT_KEY: Conventional SARIF key consumed by GitHub Code Scanning.
HEALTH_DOCS: Docs URL for the health command.
HEALTH_SCHEMA_VERSION: Current schema version for the standalone health JSON envelope.
HEALTH_SCORE_FORMULA_VERSION: Formula version for the overall health score, serialized as HealthScore::formula_version so consumers can distinguish a score shift caused by a formula change from one caused by an actual codebase change. v2 replaced the size-dependent aggregators (average and p90 cyclomatic, raw hotspot and dependency counts) with scale-invariant densities (critical-complexity share, per-thousand-file dependency rates, top-1% hotspot share) so scores are comparable across repository sizes; older snapshots that lack the density fields fall back to the v1 aggregators. See engine::vital_signs for the full penalty formula.
HOTSPOT_SCORE_THRESHOLD: Minimum churn-times-complexity hotspot score for an entry to count toward the vital-signs hotspot_count; lower-scoring entries still appear in the hotspot list but do not feed the health-score hotspot penalty.
INJECTION_NOTE: The standing injection-resistance note stamped on every guide.
JUDGMENT_ACTIONS: The closed author-action vocabulary a judgment may carry: what the receiving author (a human or the coding agent) should do with it. block and address are required actions, consider is optional, fyi needs nothing. Enforced on reentry: any other value rejects the judgment (invalid-action). The label is the reviewer’s instruction, never a fallow fact, so it stays fenced with the framing and never gates.
JUDGMENT_CONCERNS: The recommended concern vocabulary: the trade-off lenses a review reads a change through. Documented, not enforced: concern stays free text on the wire so an existing consumer keeps working, and a value from this list lets a review surface group judgments by lens.
MARKER_PREFIX_V2: Marker prefix appended to every v2 review-comment body.
MARKER_REGEX_V2: Canonical v2 marker-regex literal.
MARKER_SUFFIX_V2: Closing of the v2 marker, after the fingerprint string.
MAX_ADDED_LINES: Stop indexing added lines past this count.
MAX_COMMENT_BODY_BYTES: Hard cap on a single review-comment body, matching GitHub’s 65 536-char comment limit; bodies at or over it are truncated with a marker suffix.
MAX_CONTRACT_MEMBERS: Max contract members named inline in a coordination fact before collapsing the rest into a “+N more” suffix. Keeps the most load-bearing line readable in a terminal without discarding the trailing guidance.
MAX_DIFF_BYTES: Refuse to parse a unified diff larger than this.
MI_DENSITY_MIN_LINES: Minimum lines of code for full complexity density weight in the MI formula.
PROJECT_LEVEL_RULE_IDS: Rule ids whose findings describe project-wide config state rather than a change touching a specific source line.
PR_DECISION_SCHEMA: Schema discriminator serialized into PrDecisionSurface::schema.
PR_DETAILS_SCHEMA: Schema discriminator serialized into PrDetailsArtifact::schema.
REVIEW_BRIEF_SCHEMA_VERSION: Wire version for the fallow audit --brief --format json envelope.
RUNTIME_STALE_AFTER_DAYS: Staleness cutoff in days, mirrored from the cloud runtime-context (RUNTIME_STALE_AFTER_DAYS) so the local and cloud contracts agree.
SARIF_FINGERPRINT_KEY: Fingerprint key used in SARIF partialFingerprints and other CI formats.
SECURITY_SCHEMA_VERSION: Current fallow security --format json schema version.
SFC_TEMPLATE_SUPPRESS_COMMENT: Suppression comment recommended for single-file-component markup templates (.svelte, .vue, .astro). The // form is not markup syntax there and would render into the DOM as visible text.
SIMILAR_CODE_CACHE_CLEAR_SCHEMA_VERSION: Current vector-cache clear envelope schema version.
SIMILAR_CODE_INSPECT_SCHEMA_VERSION: Current similar-code inspect envelope schema version.
SIMILAR_CODE_REVIEW_SCHEMA_VERSION: Current similar-code review envelope schema version.
SIMILAR_CODE_SCHEMA_VERSION: Current raw similar-code envelope schema version.
SIMILAR_CODE_STATUS_SCHEMA_VERSION: Current local-provider status envelope schema version.
SNAPSHOT_SCHEMA_VERSION: Current snapshot schema version. Independent of the report’s SCHEMA_VERSION. v2: Added score and grade fields. v3: Added coverage_model field. v4: Added risk profiles (unit_size_profile, unit_interfacing_profile) and coupling concentration (p95_fan_in, coupling_high_pct). v5: Added duplication penalty to health score formula. v6: Added total_loc to vital signs (always computed from parsed modules). v7: MI formula dampening for small files (values change for files < 50 lines). v8: Added scale-invariant tail/density metrics for health score calibration. v9: Added render fan-in concentration (p95_render_fan_in, render_fan_in_high_pct, max_render_fan_in), the component-graph analogue of module fan-in / coupling concentration. Additive optional fields (matches the v4 precedent that added coupling concentration).
STYLING_HEALTH_FORMULA_VERSION: Formula version for the styling-health score (the CSS / design-system axis). Bumped independently of HEALTH_SCORE_FORMULA_VERSION whenever the styling penalty rubric is recalibrated, so consumers can distinguish a score shift caused by a weight change from one caused by an actual codebase change. v2 recalibrated dead_surface (size-stable declaration-share denominator) and token_erosion (gently saturating arbitrary-value term) from real-project evidence. v3 re-weighted the duplication family toward value DRIFT: it down-weighted the exact-block duplication scale (exact CSS duplication is the least-harmful pattern) and added a hardcoded-value-sprawl drift sub-term to token_erosion (distinct un-tokenized box-shadow/border-radius/line-height values). See engine::health::styling_score for the full rubric + calibration.
TOKEN_CONSUMER_SAMPLE_CAP: Maximum number of consumer locations sampled into TokenConsumers::consumers. The full count is preserved in TokenConsumers::consumer_count (aggregate-before-truncate), so capping the sample never distorts the count.
TYPE_AWARE_STATUS_SCHEMA_VERSION: Current schema version for type-aware status JSON output.

Statics§

CODECLIMATE_RESULT_CODES: Result issue codes emitted by the dead-code CodeClimate formatter.

Functions§

agent_schema: The default agent schema descriptor.
annotate_codeclimate_issues: Add a top-level grouped property to each typed CodeClimate issue.
annotate_sarif_results: Add a SARIF result property by resolving each result URI through a caller.
append_sarif_findings: Append SARIF findings by extracting normalized fields from typed issues.
apply_config_fixable_to_duplicate_exports: Mark every duplicate-export finding as fixable through a config edit when the project’s config supports automated fixes.
apply_path_prefix: Prefix prepended to a rendered path so CI platforms, which address files from the repository root, can find it. Empty when the analysis root already is the repository root.
apply_root_kind: Apply a document-root discriminator.
attach_telemetry_meta: Attach telemetry metadata to a JSON root object when a run id is available.
body_matches_review_id: Whether a rendered body belongs to the expected review scope.
build_audit_next_steps: Next-steps for fallow audit.
build_audit_next_steps_input: Build audit next-step inputs from typed analysis payloads plus the caller-supplied runtime suggestions gate.
build_check_output: Build the typed dead-code JSON envelope from engine results.
build_codeclimate_issue: Build a single CodeClimate issue from a stable contract descriptor.
build_combined_next_steps: Aggregated next-steps for bare fallow combined output.
build_coverage_analyze_output: Build the fallow coverage analyze --format json envelope.
build_dead_code_next_steps: Next-steps for standalone fallow dead-code.
build_dead_code_sarif: Builds the complete SARIF run value for a dead-code analysis.
build_decision_surface_output: Project a DecisionSurface into the separable, action-bearing envelope.
build_dupes_next_steps: Next-steps for standalone fallow dupes.
build_dupes_output: Build a duplication JSON envelope from caller-owned report data.
build_duplication_sarif: Build SARIF output from duplication analysis results.
build_feature_flags_output: Build the typed feature flags output envelope.
build_grouped_duplication_sarif: Build grouped SARIF output from duplication analysis results.
build_health_finding_actions: Compute the typed actions list for a complexity finding.
build_health_next_steps: Next-steps for standalone fallow health.
build_health_next_steps_input: Build standalone health next-step inputs from a typed health report plus caller-supplied runtime probes.
build_health_sarif: Build SARIF output from a health report.
build_review_brief_json_output: Build the complete fallow audit --brief --format json value.
build_sarif_document: Build a SARIF 2.1.0 document envelope.
build_sarif_result: Build a single SARIF result object.
build_sarif_result_with_snippet: Build a single SARIF result object with optional source snippet evidence.
build_sarif_rule: Build a SARIF rule object.
build_suppression_inventory_output: Build the typed suppression inventory envelope: sort by (path, line), group per file, and compute the summary (including the stale join).
cap_body_with_marker: Append marker_line to content, truncating content on a char boundary so the whole body stays within MAX_COMMENT_BODY_BYTES. The marker is never sacrificed. Returns the body and whether truncation happened.
cap_names: Cap an arbitrary list of names for inline display: first max names, then a “+N more” sentinel. Shared by the out-of-diff consumer fact in both surfaces.
check_meta: Build the _meta object for fallow dead-code --format json --explain.
clean_decision_fact: The clean, surface-agnostic fact text for a decision question, for the tour.
clone_family_actions: Build the stable action list for a clone family.
clone_group_actions: Build the stable action list for one clone group.
codeclimate_fingerprint_hash: Compute a deterministic fingerprint hash from key fields.
codeclimate_issues_to_value: Serialize typed CodeClimate issues to the wire-shape JSON array.
command_title: Human-readable report title for a fallow command name, e.g. dupes maps to “duplication report”.
composite_fingerprint: Order-independent fingerprint for a comment merged from several findings: constituents are sorted before hashing and the result carries a merged: prefix.
compute_finding_severity: Compute the severity tier for a complexity finding.
count_applied_fixes: Count fix entries whose applied flag is true.
coverage_analyze_meta: Build the _meta object for fallow coverage analyze --format json --explain.
coverage_setup_meta: Build the _meta object for fallow coverage setup --json --explain.
dead_code_docs_url: Public docs URL for a dead-code explanation anchor on the dead-code page.
decisions_outside_units: The surfaced decisions whose anchor_file is not a direction unit, in surface (rank) order. A dependency decision anchors on a package.json, which is never a graph module, so without this a dependency-only change renders as “0 files” while the JSON carries the top-ranked decision.
default_marker_regex: Default for ReviewEnvelopeOutput::marker_regex.
default_marker_regex_flags: Default for ReviewEnvelopeOutput::marker_regex_flags.
dupes_meta: Build the _meta object for fallow dupes --format json --explain.
ensure_unique_result_fingerprints: Give every result in one run its own partialFingerprints value.
escape_md: Escape a string for inclusion in a Markdown table cell.
feature_flags_meta: Metadata emitted when fallow flags --explain --format json is requested.
github_check_conclusion: GitHub check conclusion for a set of findings: Failure when any is major or above, Success when empty, Neutral otherwise.
group_review_issues_by_path_line: Group consecutive same-(path, line) issues. Input is already sorted by (path, line, fingerprint) so a single linear pass collects runs.
harmonize_dead_code_health_suppress_line_actions: Merge same-line suppress actions across dead-code and health sections.
harmonize_multi_kind_suppress_line_actions: Merge same-line suppress actions so multi-kind findings share one comment.
health_meta: Build the _meta object for fallow health --format json --explain.
health_report_from_saved_value: Rehydrate the finding-bearing portion of a saved health JSON section.
impact_digest_summary: Render the human-readable impact counter summary shared by JSON and human output surfaces.
is_false: Helper for skip_serializing_if = "is_false" on truncated fields.
is_judgment_action: Whether action is one of JUDGMENT_ACTIONS.
is_project_level_rule: Whether findings for rule_id describe the whole project (e.g. dependency rules) rather than a specific file location.
issue_output_contract_by_code: Output-facing dead-code result contract by issue code.
issue_output_contracts: Output-facing dead-code result contracts in stable registry order.
issues_from_codeclimate: Extract normalized CI issues from a raw CodeClimate JSON array, sorted by severity then location.
issues_from_codeclimate_issues: Normalize typed CodeClimate issues into CI issues, sorted by severity then location.
letter_grade: Map a numeric score (0-100) to a letter grade.
markdown_code_span: Render a complete CommonMark code span around an untrusted value. The fence grows past the longest backtick run inside the value, so the span cannot be closed early from within.
markdown_table_code_span: markdown_code_span for a Markdown table cell: pipes are additionally escaped so the value cannot terminate the cell, and line endings collapse to spaces because any CommonMark line ending would split the table row.
markdown_table_text: Escape prose for a Markdown table cell while leaving intentional inline markup alone: pipes are escaped and line endings collapse to spaces.
normalize_uri: Normalize a path string to a valid URI: forward slashes and percent-encoded brackets.
parse_new_hunk_start: Parse the new-file start line out of a @@ -a,b +c,d @@ hunk header.
parse_review_id_marker: Parse the sole canonical review-scope marker from a rendered body.
plan_pr_comment_post: Decide whether the rendered comment should be created, updated, or skipped.
pr_status_contexts: Per-gate status contexts for surface; shorthand for pr_status_contexts_with_mode with PrStatusMode::Split.
pr_status_contexts_with_mode: Commit-status contexts for surface, fanned out according to mode.
relative_to_diff_path: Relativize path against root with forward slashes, passing an already relative path through unchanged. None for absolute paths outside root.
render_pr_comment: Render the sticky summary comment body: identity marker, run verdict, headline count, and per-category findings tables.
render_pr_comment_with_verdict: Render the sticky summary comment body with an explicit gate conclusion.
render_pr_summary: Renders the sticky PR summary comment body, prefixed with its identity marker and carrying finding-count truncation metadata in the envelope.
render_review_comment_for_group: Render one comment from a group of issues sharing the same (path, line).
render_review_envelope: Render a provider-specific review envelope from typed CI issues.
render_review_envelope_with_conclusion: Render a review envelope with an explicit gate conclusion and status.
render_scoped_review_envelope: Render a review envelope whose bodies carry the supplied review scope.
render_scoped_review_envelope_with_conclusion: Render a scoped review envelope with an explicit gate conclusion and status.
review_id_marker: Render the exact review-scope marker line.
review_label_from_codeclimate: Map a CodeClimate severity name to the review badge label: error for major and above, warn otherwise.
rule_docs_url: Public docs URL for a rule’s docs_path relative to the docs site root.
sarif_finding_fingerprint: Stable SARIF fingerprint for a finding with source snippet evidence.
security_meta: Build the _meta object for fallow security --format json --explain.
serialize_audit_json_output: Serialize a typed fallow audit --format json envelope with the standard root discriminator policy.
serialize_check_grouped_json_output: Serialize fallow dead-code --group-by ... --format json.
serialize_check_json_output: Serialize fallow dead-code --format json.
serialize_combined_json_output: Serialize a typed bare fallow --format json combined envelope with the standard root discriminator policy.
serialize_coverage_analyze_json_output: Serialize the fallow coverage analyze --format json envelope.
serialize_coverage_setup_json_output: Serialize the fallow coverage setup --json envelope.
serialize_cross_repo_impact_json_output: Serialize the fallow impact --all --format json envelope.
serialize_decision_surface_json_output: Serialize the standalone decision-surface envelope.
serialize_doctor_json_output: Serialize a doctor report with its root discriminator.
serialize_dupes_json_output: Serialize fallow dupes --format json.
serialize_explain_json_output: Serialize the fallow explain --format json envelope.
serialize_feature_flags_json_output: Serialize fallow flags --format json.
serialize_fix_json_output: Serialize the typed fix JSON root.
serialize_health_json_output: Build and serialize a health root JSON envelope.
serialize_impact_json_output: Serialize the fallow impact --format json envelope.
serialize_inspect_json_output: Serialize the fallow inspect --format json envelope.
serialize_json_root_output: Serialize a typed fallow root envelope with the requested discriminator mode.
serialize_list_boundaries_json_output: Serialize fallow list --boundaries --format json.
serialize_list_workspaces_json_output: Serialize fallow list --workspaces --format json.
serialize_named_json_output: Serialize an output envelope and apply an explicit root discriminator.
serialize_review_brief_json_output: Serialize the fallow audit --brief --format json envelope.
serialize_review_envelope_json_output: Serialize the review envelope contract emitted by CI review formats.
serialize_review_reconcile_json_output: Serialize the review reconcile contract.
serialize_scoped_review_envelope_json_output: Serialize a scoped review envelope through the canonical derived wire type.
serialize_security_blind_spots_json_output: Serialize the fallow security blind-spots --format json envelope.
serialize_security_json_output: Serialize the full fallow security --format json envelope.
serialize_security_summary_json_output: Serialize the compact fallow security --summary --format json envelope.
serialize_security_survivors_json_output: Serialize the fallow security survivors --format json envelope.
serialize_similar_code_cache_clear_json_output: Serialize a vector-cache clear result with its root discriminator.
serialize_similar_code_inspect_json_output: Serialize a similar-code inspect packet with its root discriminator.
serialize_similar_code_json_output: Serialize raw similar-code output with its root discriminator.
serialize_similar_code_review_json_output: Serialize reviewed similar-code output with its root discriminator.
serialize_similar_code_status_json_output: Serialize local-provider status with its root discriminator.
serialize_suppression_inventory_json_output: Serialize fallow suppressions --format json.
serialize_trace_error_json_output: Serialize the fallow trace-error --format json envelope.
serialize_trace_json_output: Serialize the fallow trace --format json envelope.
serialize_type_aware_status_json_output: Serialize the type-aware status envelope with its root discriminator.
serialize_walkthrough_guide_json_output: Serialize the review walkthrough guide envelope.
serialize_walkthrough_validation_json_output: Serialize the review walkthrough validation envelope.
strip_root_prefix: Recursively strip a project-root prefix from all string values in a JSON tree.
styling_audit_context_label: Renders the parenthesized gate-context label for a styling audit finding.
styling_candidate_count: Sums every descriptive CSS finding count in summary, saturating at u32::MAX.
summarize_coverage_source_consistency: Summarise the coverage-source provenance attached to CRAP findings.
summary_fingerprint: Stable fingerprint for a summary comment body.
summary_label: Collapsible-section label for a findings category: appends “showing N” when the table is capped below the category’s total.
suppress_comment: Build the suppression comment a decision’s suppress action pastes in.
trace_unused_export_input: Select the deterministic unused-export target used by read-only trace next-step commands.
validate_review_body_scope: Require a body marker to match the expected review scope exactly.
validate_saved_security_envelope: Validate a saved security envelope against the current output-owned schema.
visible_stage_units: The visible stage members for a guide: direction units in order, MINUS any file collapsed into Cleared (de-prioritized or viewed). Returned as references into the guide so the caller can render rows without cloning.

Type Aliases§

BranchingSnapshot: Branching totals per root-relative path, for one revision.
StandardReviewBriefOutput: The standard audit brief payload shape used by the CLI, schema emitter, API, and agent-facing review surfaces.
StandardWalkthroughGuide: The standard walkthrough guide shape emitted by fallow review.

Crate fallow_output

Crate fallow_output Copy item path

Modules§

Structs§

Enums§

Constants§

Statics§

Functions§

Type Aliases§

Crate fallow_output