Imported from avsngh-git/RAGpipeline (
AGENTS.md). Install upstream withnpx skills add avsngh-git/RAGpipeline. Copyright stays with the author.
Project guidance
Before planning, writing code, changing architecture, selecting infrastructure,
or adding dependencies, read
docs/agents/scientific-research-platform-source-of-truth.md in full.
That document is the authoritative project specification and high-level plan. Use its delivery phases and gates to scope work, its LOCKED and OPEN decisions to guide choices, and its change-control process for material changes. Keep the specification at its current path as the single source of truth.
Before making a material contribution, apply its Agent operating checklist. Choose the smallest coherent implementation that serves the current phase.
Agent skills
Learning workflow
For Phase 0 work, read docs/plans/phase-0-learning-handoff.md.
For Phase 1 work, read docs/plans/phase-1-learning-handoff.md, then follow the
approved tasks in docs/plans/phase-1-corpus-ingestion.md.
These handoffs record progress and the next exercise. The user implements the
project to learn Python; default to tutoring and reviewing, and implement only
when explicitly asked.
Issue tracker
Track issues and specs in GitHub Issues for avsngh-git/RAGpipeline.
Before reading or publishing tickets, read docs/agents/issue-tracker.md.
Triage labels
Use the five default triage labels.
Before triaging issues, read docs/agents/triage-labels.md.
Domain docs
Use a single-context layout: root CONTEXT.md and docs/adr/.
Before exploring domain concepts or architectural decisions, read
docs/agents/domain.md.