Archive to Live Route Reconciliation
Turn Wayback, Common Crawl, and public urlscan artifacts into a source-linked route delta, then promote only currently reachable in-scope behavior.
Enter with artifacts, leave with evidence.
- archive-only endpoint
- old JavaScript route
- renamed API
- historical parameter
- post-migration review
- dated scope snapshot
- public archive use allowed
- low-rate live verification budget
Agents can search this workflow, retrieve the complete graph, or request one stage through the read-only Vulns.co MCP connector.
Snapshot each archive source
Cache bounded Wayback CDX, Common Crawl CDXJ, and public urlscan search results with exact query, index, filters, and retrieval time.
- approved roots
- query manifest
- immutable source snapshots
- retrieval manifest
- Evidence gate
- Every historical row retains source, source timestamp, original URL, and raw snapshot reference.
- Negative control
- A known off-root or excluded name is rejected during normalization.
- Stop condition
- Do not submit private targets as public scans or query outside approved roots.
Cluster route families
Normalize host case and fragments, then group paths, parameters, extensions, scripts, and operation nouns without losing original values.
- source snapshots
- route families
- parameter vocabulary
- script inventory
- Evidence gate
- Every normalized member links back to at least one exact source row.
- Negative control
- Third-party and shared-service URLs remain separate from target-owned candidates.
- Stop condition
- Do not replay archived request bodies, tokens, or personal data.
Compare current behavior
Probe only current in-scope candidate URLs at a bounded rate and retain status, redirect, content type, body length, title, and timestamp.
- prioritized candidate URLs
- current known route
- random nonexistent route
- current response ledger
- archive-to-live state labels
- Evidence gate
- A promoted URL has a current response plus both known-route and nonexistent-route controls.
- Negative control
- Uniform CDN, wildcard, soft-404, and redirect behavior is identified before a route is called reachable.
- Stop condition
- No state-changing methods, archived POST bodies, or broad replay.
Promote a falsifiable lead
Rank independent-source agreement, route sensitivity, current behavior, and safe test cost, then assign one discriminator.
- route delta
- control results
- evidence-backed lead
- decay date
- next discriminator
- Evidence gate
- The lead separates historical fact, current observation, inference, and untested security claim.
- Negative control
- Archive-only and dead routes remain historical notes rather than active leads.
- Stop condition
- Stop when current behavior cannot distinguish the candidate from the control.