IncidentSutra AI documentation
Observe precisely. Reason cautiously. Recover responsibly. Created by Vatsal Shah. Apache-2.0.
This folder is the complete reference for IncidentSutra AI v0.2.1. Every page exists as Markdown (source) and HTML (generated, styled, light and dark, print-ready). Rebuild HTML with npm run build:docs; CI fails if it is stale.
Read in this order
| Audience | Start here |
|---|---|
| Executives, managers, procurement | 01 Executive summary |
| Product and incident commanders | 02 Functional specification |
| Engineers and contributors | 03 Technical specification, 04 Architecture |
| On-call users | 05 User guide |
| Security, privacy, legal, compliance reviewers | 06 Security, privacy and legal |
| Anyone deciding whether to trust it | 07 Independent audit and evidence, Test results |
| People using AI coding tools | 08 Agent catalogue |
| Quick answers | 09 FAQ |
Evidence folder
evidence/ holds machine-written artifacts: test-results.json, benchmark.json, audit/audit-report.{html,md,json}, and samples/ (a full synthetic incident report in HTML, Markdown and JSON). None of it is edited by hand. Regenerate with npm run test:report, npx incidentsutra audit --out ..., npx incidentsutra benchmark --out ... and npm run demo.
Ground rules this documentation follows
- Claims are limited to what the code does and a test or audit check demonstrates.
- Anything not built is labelled planned, never implied.
- The project is advisory tooling. It does not replace a human incident commander, and nothing here is legal, compliance or security certification.