Ingestion Review Checklist

Use this checklist to review artifacts produced by /sniper-ingest.

Project Brief (`docs/brief.md`)

[ ] Accuracy: Brief accurately describes what the project does (not aspirational or hallucinated)
[ ] Problem statement: Clearly identifies the problem the project solves
[ ] Target users: Identifies who uses the project (inferred from code, not guessed)
[ ] Current scope: Lists features that actually exist in the codebase
[ ] Constraints: Notes technologies and external dependencies that are locked in
[ ] Honest gaps: Sections that couldn't be inferred are marked "Unable to determine"

[ ] Component diagram: Includes an ASCII or Mermaid diagram that matches actual code structure
[ ] Directory alignment: Component boundaries match the actual directory structure
[ ] Data models: Extracted data models match actual ORM/migration/schema files
[ ] API contracts: Documented endpoints match actual route definitions
[ ] Technology choices: Listed technologies match actual dependencies (package.json, etc.)
[ ] Infrastructure: Infrastructure description matches Docker/Terraform/K8s files (if present)
[ ] Cross-cutting concerns: Auth, logging, error handling patterns are documented
[ ] No hallucinations: No components, APIs, or data models that don't exist in the code

[ ] Code examples: Every convention cites a real code example with file path
[ ] Representative sampling: Conventions are derived from multiple files, not just one
[ ] Inconsistencies noted: Where patterns conflict, both are documented
[ ] Complete coverage: Naming, code org, error handling, testing, API, imports, config, logging
[ ] Conventions vs accidents: Patterns appear in 80%+ of relevant files to qualify as conventions

[ ] Ownership paths updated: The ownership section in .sniper/config.yaml reflects the actual project directory structure, not the template defaults
[ ] All directories covered: Major source directories are assigned to an ownership category

[ ] Consistency: Brief, architecture, and conventions don't contradict each other
[ ] Sufficient depth: Artifacts provide enough context for agents to work on the codebase
[ ] Actionable: A developer or agent reading only these artifacts could understand the project's structure and patterns