build-incident-history

You can build an incident history archive by feeding raw Slack exports, GitHub artifacts, and Tsuga CLI logs into a processor that generates validated SUMMARY.md dossiers and retrieval indexes. This transforms heterogeneous incident inputs into structured, searchable records.

Structuring incident metadata for retrieval requires generating metadata.json fields and an _inventory.csv index alongside per-incident SUMMARY.md files. This deterministic structure with canonical section order enables downstream agents to reliably search and reason over incident histories.

Yes, you can bootstrap an incident knowledge base from GitHub PR artifacts, incident reports, and Tsuga CLI command logs. The process accepts these heterogeneous sources to reconstruct incident narratives and output validated dossiers under an incidents directory.

Preserving a diagnostic path in incident summaries requires a transformation process that enforces canonical section order and documents real tsuga CLI syntax directly within a validated SUMMARY.md dossier for each incident.

Yes, building the archive requires placing raw incident inputs like metadata, timelines, and paging context into local inputs/incidents/<INC-id>/ folders so the processor can generate the corresponding output dossier and inventory files.

A structured incident archive replaces raw dumps with validated SUMMARY.md dossiers and metadata-driven artifacts, solving the problem of unreliable search by providing deterministic structure and stable retrieval indexes for downstream reasoning.