This is the primary entry point for any AI agent working in this codebase. It provides a categorized, status-tagged index of all 86+ documentation files, a task-based reading order, and a status taxonomy so you can immediately distinguish current architecture from historical fix notes from unimplemented plans.
Do not read all 86 files. Use the task-based reading order below to find exactly what you need, then follow the category dispatch table to the relevant index.
Every document in every category index is tagged with one of four statuses:
| Status | Meaning |
|---|---|
current |
Describes the system as it exists today. Trust it fully — code matches. |
legacy-canonical |
Describes an older version or subsystem that must remain compatible. Trust it for the version it describes. Critical for migration and format work — do not dismiss it as outdated. |
historical |
A problem was diagnosed and fixed; the code already incorporates the solution. Read for context when working in that area. Do not copy configuration snippets or line numbers verbatim — the code has moved on. |
future-intent |
An architectural plan or vision document. Design decisions are recorded here but no implementation exists yet. Do not assume any code matches. |
deprecated is intentionally absent from this taxonomy. If a document is truly obsolete, its index entry carries a superseded-by: note pointing to what replaced it.
Find your task below and read only the listed files — in order. This is the fastest path to being productive.
| Task | Files to read (in order) |
|---|---|
| Understand core architecture | CLAUDE.md, README.md (§Architecture section) |
| Work with the Wizard / MCP | AI_INTEGRATION_GUIDE.md, REDSTRING_MCP_SYSTEM_PROMPT.md, then AGENTIC_ARCHITECTURE.md (historical — explains the pipeline shape) |
| Call MCP tools from an external client | MCP_TOOLS_QUICK_REFERENCE.md, MCP_SETUP_GUIDE.md |
| Use Redstring headless (no browser: CLI, workspaces, universes, GitHub) | HEADLESS.md (current — the CLI, workspace/universe model, redstring init, pull/push) |
Read or write .redstring files |
redstring-format-spec.md (legacy-canonical — migration code is derived from this), REDSTRING_FORMAT_VERSIONING.md |
| Migrate data between format versions | REDSTRING_FORMAT_VERSIONING.md, MIGRATION_GUIDE.md, redstring-format-spec.md (legacy-canonical) |
| Modify the layout algorithm | AUTO_LAYOUT_GUIDE.md, REDESIGNED_LAYOUT_SUMMARY.md (historical — explains why it was rebuilt), FORCE_SIMULATION_TUNER.md |
| Investigate a drag or performance regression | DRAG_PERFORMANCE_COMPLETE.md (historical — contains exact NodeCanvas/utils.js line references from the three-bottleneck analysis) |
| Deploy or configure infrastructure | DEPLOYMENT.md, GITHUB_APP_SETUP.md, cloudflare/README.md (if targeting Cloudflare Pages) |
| Work with SPARQL / semantic web / Wikidata | SEMANTIC_WEB_INTEGRATION.md, RDF_INTEGRATION_README.md, SEMANTIC_DISCOVERY_GUIDE.md |
| Set up the save / sync system | SAVE_COORDINATOR_README.md, GIT_FEDERATION.md |
| Understand the v4.0.0 format roadmap | FORMAT_REFACTOR_PLAN.md (future-intent — SKOS/PROV/RDF-star alignment; no code exists yet) |
| Run or write tests | TESTING_ONBOARDING.md, AI_TESTING_GUIDE.md, WIZARD_TESTING_GUIDE.md |
| Understand project philosophy / conceptual vocabulary | aiinstructions.txt (plain text, not .md — read via cat aiinstructions.txt) |
| Category | Index file | Read when |
|---|---|---|
| Core System Reference | .compendium/core-system.index.md |
Starting any task; foundational architecture |
| AI Agent and MCP | .compendium/ai-agent-mcp.index.md |
Working with Wizard, MCP server, bridge daemon, prompts |
| Data Format and Migration | .compendium/data-format.index.md |
Reading/writing .redstring; format versioning; migration logic |
| Graph Engine and Layout | .compendium/graph-layout.index.md |
Force simulation, layout algorithms, constraint systems |
| Storage, Sync, and Federation | .compendium/storage-federation.index.md |
Universe management, Git integration, deployment, OAuth |
| Semantic Web and Knowledge Discovery | .compendium/semantic-web.index.md |
RDF, SPARQL, Wikidata/DBpedia, semantic enrichment |
| Development Operations | .compendium/dev-ops.index.md |
Refactoring, testing, audits, UI fixes, performance ops |
- Last reviewed: 2026-06-13
- File count indexed: 86 .md files +
aiinstructions.txt - Files intentionally excluded:
docs/README.md(Mintlify template placeholder — contains no Redstring content) - No existing files were moved or renamed — all index entries point to files at their original paths
- When adding a new significant .md file to the project, add an entry to the relevant
.compendium/*.index.mdand update this file's task table if warranted