Find unused code.
See the evidence.
Run your analyzers. Review every finding in one clear report.
Local retained history
Repositories
Runs kept by this sift process (bounded retention). Not a permanent repository database.
Loading local run history…
Supported integrations
Analyzers
Engine selection for the next run is on Analyze. Per-run task results appear in Report → Analyzer runs.
No run selected
Report
Analyze a public repository or open a retained run from Repositories.
Go to AnalyzeNo source was changed. Check that the repository is public and that analyzer tools (fallow, knip, vulture) are available on this host.
| Select | Finding | Location | Classification | Removal |
|---|
Runtime coverage
Attach coverage to this run
Workspace applies to runtime evidence identity only — not static scan scope. Missing runtime evidence is not proof of deadness.
Open a completed static analysis run first.
Unified report
Static findings
Observation-based · review only
Runtime findings
Analyzer task results appear after a run produces a report.
Draft removal workflow
Record, authorize, verify, publish
Select eligible findings in Findings, then complete the four steps below. Classification and baseline are warnings, not selection locks.
Checking draft publisher status…
Creates human dispositions for newly selected findings only (already-recorded findings are skipped).
Authorizes exact disposition IDs returned by step 1 (or already on this run).
Reacquires the commit and builds an integrity-checked aggregate patch for authorized findings.
Publishes a draft pull request only. Never merges or marks ready.
Advanced · combined create-pr (API compatibility)
Runs dispose → authorize → aggregate → publish in one request. Prefer the four steps above.
Verified patch artifact
No patch is available yet.