Static question
Ask: which selected endpoint, client call, route, root method, or source root is being inspected?
Stop condition: ambiguous endpoint/root, private-only selector, or unsupported selector.
Route-flow proof path
Follow the selected static route trail, then stop where evidence stops.
This concept page explains how a reader can inspect route-centered static evidence from a selector or root toward route/root evidence, bridge state, selected rows, context rows, gaps, limitations, and owner follow-up.
Public claim level: concept. No public conclusion without evidence. This page is a reading model, not a public demo result, runtime trace, scanner guarantee, reducer conclusion, approval flow, or operational certification.
Placement
This is one proof-path family, not the whole proof model.
/proof-paths/Broader proof-path overview; this page focuses on the route-centered family.
/proof-paths/tour/Guided reading flow; this page names route-flow row and stop vocabulary.
/evidence/General evidence vocabulary; this page applies those fields to route-flow rows.
/limitations/Canonical boundary surface; this page carries limitations beside each route-flow claim.
/static-vs-runtime/Static evidence versus telemetry boundary; this page keeps route-flow classifications static.
/review-claim-checklist/Repeat, downgrade, hold, or keep-internal ritual; this page helps inspect the route-flow evidence before that ritual.
/glossary/Canonical term index; this page uses those terms in a route-flow context.
Proof path anatomy
Every public row keeps the evidence fields attached.
Evidence path
Follow: selector/root to route/root evidence, bridge state, selected static flow row, context group, gap, and owner follow-up.
Stop condition: missing proof path or unjoined adjacent context.
Rule ID or rule family
Record: examples include combined.route-flow.selector.v1, combined.route-flow.path.v1, combined.route-flow.gap.v1, and the broader combined.route-flow.* family.
Stop condition: missing rule ID or missing documented rule family.
Evidence tier
Compare: Tier1Semantic, Tier2Structural, Tier3SyntaxOrTextual, or Tier4Unknown.
Stop condition: missing evidence tier, or syntax/textual evidence repeated as semantic proof.
Coverage label
Label: full, partial, reduced, unknown, unavailable, future-only, or gap-labeled coverage as a boundary.
Stop condition: missing coverage label or reduced coverage affecting the selected path.
Supporting IDs
Preserve: public-safe row IDs, fact IDs, edge IDs, symbol IDs, source labels, and artifact IDs when available.
Stop condition: supporting IDs are raw private identifiers, local artifact paths, or hidden detail.
Public-safe source context
Record: file path and line-span style context, snippet hashes, commit/source context, and extractor version or schema family when public-safe.
Stop condition: raw snippets, raw values, missing source context, or missing extractor/schema context.
Limitation or non-claim
Hold: every row carries a limitation, or a documented closed-set equivalent saying no additional row limitation is present.
Stop condition: missing limitation or unsupported demo claim.
Next owner/question
Hand off: route reviewer, implementation owner, validation owner, or service owner gets the next question when evidence stops.
Stop condition: no owner or next question after a gap.
Current-branch evidence rule
Checked-in route-flow evidence supports vocabulary, not broad public completion.
Current checked-in code and rule catalog entries define the route-flow report shape, rule IDs, classifications, context groups, selected projection rows, and gap rows. This page does not turn those source facts into a public demo claim.
Backed now
route-flow-report.md and route-flow-report.json are checked-in report names in code and rule catalog entries, with deterministic route-flow rule families and static classifications.Backed nowRule families such as
combined.route-flow.entry.v1, combined.route-flow.dependency-surface.v1, combined.route-flow.fact-symbol-projection.v1, combined.route-flow.classification.v1, and combined.route-flow.redaction.v1 are documented in the rule catalog.Limited nowAttachment precision has multiple checked-in slices, while broader families remain in-progress or future-only unless cited by narrower public-safe evidence.
Deferred publiclyReal route values, local generated outputs, private repository labels, and raw scan internals stay out of public copy.
Row vocabulary
Selected context must join through the proof path.
Selected rows
Allowed row labels include selector, endpoint/root,
route/root evidence, bridge state,
static flow row, service/helper,
repository/data, query or SQL shape,
dependency surface, value origin, and
implementation candidate.
A row is selected only when it joins through selected route-flow evidence, selected source-local symbol identity, selected argument/parameter evidence, or another documented rule-backed relationship.
Context and gaps
context group, gap, limitation,
and owner follow-up rows keep useful review context
visible without promoting adjacent evidence into a selected path.
Adjacent unjoined service, data, query, dependency, or value evidence is a gap or context limitation, not a selected route-flow path.
Static classifications
Route-flow classification labels are static only:
StrongStaticRouteFlow, ProbableStaticRouteFlow,
NeedsReviewStaticRouteFlow, NoRouteFlowEvidence,
and UnknownAnalysisGap.
NeedsReviewStaticRouteFlow and
UnknownAnalysisGap are review states, not failures,
quality judgments, or owner assignments.
Static boundary
The strongest static label is still not runtime proof.
StrongStaticRouteFlowis still static evidence; it does not prove runtime request execution, runtime binding, production traffic, endpoint performance, release safety, outage cause, business impact, or complete coverage.- Route-flow rows do not prove runtime dependency-injection target selection, branch feasibility, SQL execution, database state, or data contents.
- Route-flow is not AI impact analysis, LLM analysis, embeddings, vector databases, prompt classification, autonomous approval, release approval, or a replacement for tests, source review, human review, runtime observability, or service-owner judgment.
- Public copy must not publish raw source, raw SQL, raw config, secrets, raw local paths, raw repository remotes, private sample names, private route values, generated output directories, raw facts, raw SQLite, analyzer logs, command output, hidden validation detail, or credential-like values.
Review language
Use bounded outcomes and verbs.
Safe verbsinspect, follow, compare, record, label, downgrade, hold, hand off, and escalate.
Review outcomes
show as static evidence, show as context, label the gap, downgrade, keep internal, owner follow-up, and do not repeat.Illustrative safe patternThis route-flow row shows static evidence under Rule
combined.route-flow.* with Tier2Structural, partial coverage, supporting IDs, and limitations. Illustrative only; not a real TraceMap finding.Illustrative gap patternThis adjacent context is not joined to the selected route-flow path, so the public story labels a gap instead of implying a flow. Illustrative only.
Illustrative classification patternThe classification is static; runtime behavior and release decisions remain outside this evidence story. Illustrative only.
Rejected patterns
Do not use these as live claims.
- Rejected pattern: route-flow proves runtime behavior.
- Rejected pattern: route-flow proves production traffic or endpoint performance.
- Rejected pattern: route-flow identifies outage root cause.
- Rejected pattern: route-flow proves the release is safe.
- Rejected pattern: route-flow proves complete coverage.
- Rejected pattern: route-flow guarantees, certifies, approves, replaces, resolves, is production-proven, is safe to release, or gives unqualified impact wording.
- Rejected pattern: route-flow is AI impact analysis or LLM impact analysis.
Stop conditions
Stop, downgrade, or hand off when the proof path breaks.
- Missing proof path, missing rule ID or rule family, missing evidence tier, missing coverage label, missing limitation, or missing supporting public-safe source context.
- Private-only evidence, hidden detail, unjoined adjacent context, ambiguous endpoint/root, runtime-only binding, or reduced coverage that affects the selected path.
- Schema/extractor gap, unrecognized or unlisted classification label, unsupported demo claim, or forbidden runtime, release, operational, complete-coverage, business-impact, AI, or LLM wording.
Continue