Remove interpreter/serve-era staleness across the wiki: no interpreter.rs/serve.rs exist, no nirdosha serve subcommand, no --format=json flag, no live/served HTTP API; correct compiled-vs-not-currently-running claims per struct/spawn/chan/identity now being compiled while sandbox/db/json/http/mq/transact/workflow/live-UI remain not running in any form; fix stale example file paths
Fix repo-owner/reorg'd links (arunsoman -> kannamma-labs, top-level docs -> docs/, compiler -> crates/compiler); document rfcs/0009's chart grammar + UI-plugin component extensibility on UI-Engine and Honest-Scope-and-Roadmap
Build out the full wiki: design philosophy, architecture, LLM integration
Moves the deep-dive content out of the README into 12 wiki pages: Design Philosophy, Who It's For, Nirdosha vs. Alternatives, Architecture, Language Features, the UI Engine, Benchmarks, LLM Integration, Getting Started, Honest Scope & Roadmap, and FAQ, plus a Home landing page and sidebar nav.
Every page is grounded in the repo's existing docs (goal.md, compiler/src/INDEX.md, nirdosha-agent-api.md, benchmarks/RESULTS.md) rather than invented, and each claim points at the specific source file, test, or independent tool backing it. LLM Integration is written specifically for an LLM audience: what each mechanism solves for an agent, its shipped/planned status, and the evidence (generated apps, GBNF cross-checks, benchmark methodology) behind the claim.