Current Project Status
Last updated: 2026-11-17 (Content Studio track Y0–Y10 closed, plus the
daily productivity & keynote studio suite — native pure-Rust OOXML .pptx
writer, 16:9 widescreen Typst .pdf export, art-directed illustration
node, 15 Studio catalogue workflows, and Pingora production gateway; see
RoadMap.md and TODOs.md)
This is the current authoritative status entrypoint for Yanshi. Historical evaluations, roadmap notes, and TODO queues may explain how the project arrived here, but this document is the shortest maintained summary of what exists now and what remains active.
Summary
Yanshi is a multimodal content-production runtime: sources in (idea / URL /
files / recording), deliverables out (report, mind map, audio overview, slides,
flashcards, quiz, infographic, data table, video overview, domain engineering
diagrams). The contracts are in RFC_CONTENT_STUDIO.md;
the evaluation behind the reset is
PROJECT_EVALUATION_2026-11-04.md. It is
not a coding-agent framework.
Underneath it is a Rust workspace for deterministic DAG workflows, agent-native runtime loops, Skills, MCP tools, RAG, memory, tracing, plugins, distributed worker foundations, and a Web UI run console.
The current architecture is organized into five layers, with a narrow-waist
contract kernel (L0) extracted by the P-A track so the runtimes depend only on
shared contracts (enforced by cargo xtask check-arch; see
docs/RFC_CRATE_ARCHITECTURE.md and docs/ARCHITECTURE.md):
- L0 contract kernel:
yanshi-value(FlowValue),yanshi-graph(theFlowIR),yanshi-store-spi,yanshi-agent-spi,yanshi-async-util, andyanshi-tool(theToolcontract, split out ofyanshi-toolsin T3.3). - L1 execution core (the executor):
yanshi-coreruns the L0FlowIR via theFlowExttrait (flow.run()). - L2 capability adapters:
yanshi-nodes(tool-tier + deterministic renderers includingrender_slides, pure-Rustpptx_writer,render_video,render_audio,render_mindmap,render_study,render_table,render_svg,render_markdown,render_document, plusweb_search,repo_snapshot,pdf_extract,video_ingest),yanshi-nodes-ai(capability-backed node adapters:llm,figure,illustration,asr,tts,text_to_image,image_to_image,image_edit,image_understand,text_to_video,mcp,rag),yanshi-llm,yanshi-tools(builtin tool implementations + OS-sandbox backends),yanshi-tools-ai(TTS / ASR / image / video asTools),yanshi-mcp,yanshi-rag,yanshi-memory. - L3 agent and orchestration:
yanshi-agents,yanshi-skills,yanshi-harness,yanshi-config(shared config-first workflow assembly- diagnostics, consumed by both the CLI and server),
yanshi-cli.
- diagnostics, consumed by both the CLI and server),
- L4 operations and productization:
yanshi-tracing,yanshi-server,yanshi-db,yanshi-worker,yanshi-ui, plus the standalone Cloudflare Pingora TLS/HTTP2 reverse proxy incrates/gateway.
Implemented Surfaces
- DAG workflow execution through
yanshi-core::Flow(run via theFlowExttrait). - Config-first workflow validation and execution through
yanshi-cli. - Agent-native runtimes through
AgentRuntime, ReAct, Plan-Execute, reflection, memory, and supervisor patterns. - Dynamic workflow:
yanshi_agents::dynamic::compile_plan_to_flowcompiles a declarativeWorkflowPlaninto a parallelFlowof tool calls, andDynamicWorkflowAgentmakes the LLM planning call then compiles + executes. Exposed on the CLI asyanshi workflow dynamic --goal ... --model ..., where the LLM-authored plan runs against a restrictive built-in tool sandbox (--allow-path/--allow-domaingrant access; shell is never registered),--dry-runprints the plan without executing, and--approveroutes every tool call through the Harness approval pipeline. - Harness governance shell (
yanshi-harness): hooks, interactive approval, sandbox, audit, run limits, background tasks, and theHarnessEventenvelope. - Skills through
SKILL.mdandskill.toml, including tiered[[knowledge]]retrieval (inlinefilestier or indexedragtier) with pluggable chunking strategies — fixed-size, sentence, recursive, paragraph, heading, code-AST, and embedding-basedsemanticchunking (requiresOPENAI_API_KEY). - Tool abstraction through
Tool,ToolRegistry, policy, permissions, and typed output parts. - MCP client, server scaffolding, workflow nodes, CLI calls, and Skill tool integration.
- RAG search/index/eval foundations behind the
ragfeature. - Trace persistence, replay, TUI, redaction, and OpenTelemetry mapping.
- Server run APIs, event history, SSE streaming, cancellation, and embedded Web UI run console.
- Subprocess plugin runtime,
plugin.toml, workflow plugin nodes, plugin CLI, and marketplace schema support. - Distributed scheduler foundation, gRPC worker protocol, worker runtime, and stitched worker trace events.
- Official offline-first ecosystem samples under
examples/ecosystem/. - Sandbox and code-execution hardening (the S track, closed 2026-07-24
through 2026-07-27): file+script combination-chain and skill-script
integrity fixes; a real Linux 6.12 VM (via Apple’s
containerCLI) used to compile- and kernel-enforce the Landlock + cgroup v2 backend; OS sandbox now defaults on (security.os_sandbox = true);code_exec— a mandatory, strongly-isolatedContainerBackendtool for running LLM-generated Python, separate from the OS-sandbox tier, with zero network access, hardcoded resource limits, and automatic Harness approval escalation (ToolIdempotency::NonIdempotent). Seedocs/RFC_LLM_CODE_EXECUTION.md. - Long-horizon tasks and retrieval hardening (the L track, closed
2026-07-2x): replan-loop closure for stalled/failed plan steps, task-summary
recovery so a resumed session doesn’t re-derive prior progress from scratch,
project-level memory persisted across sessions, RAG retrieval
strengthening, and a delegation contract (
yanshi-agent-spi::delegationaggregation) for sub-agent hand-off with schema-validated answers and conflict-flagged result aggregation.
LLM providers
The full per-provider capability matrix, ProviderRequest contract,
ToolChoice modes, ModelCapabilities flags, model families /
context windows, and rate-limit handling all live in
LLM_PROVIDERS_MATRIX.md. That document is
the single source of truth for what each provider supports; entries
are verified by yanshi-llm/tests/provider_consistency.rs (offline)
and provider_consistency_live.rs (opt-in live).
Stability
The v1 stability inventory lives in:
These documents define stable, beta, experimental, and internal surfaces for Rust traits, manifests, trace schemas, server envelopes, and plugin/marketplace contracts.
Active Work
The short-term execution queue in TODOs.md (maintainer-local, gitignored) is
currently empty — every segment through X (2026-10-07 evaluation fixes)
and Y0–Y10 (the Content Studio reset specified by RFC_CONTENT_STUDIO.md)
has closed:
- Y0–Y2 —
RFC_CONTENT_STUDIO.md, first-classFlowValue::Fileartifacts with DB registration +/v1/runs/{id}/artifactsstreaming download + CLI--output-dir, streaming tokenusage, modalityCostMeter+cost_limit_usd, and DAG checkpoint/resume across CLI and server. - Y3 — embedded pure-Rust renderers (
render_table,render_svg,render_markdown,render_mindmap,render_study,render_slides+ native OOXMLpptx_writer), sandboxed media/document renderers (render_audio,render_video,render_documentover pandoc + typst),figurenode (svg/d2/tikz/typst/pythoncontainer with domain checks + compile-error feedback loop), andillustrationnode (art-directed prompt refinement, 6 visual style presets, dark/light palette harmonization, bounded concurrentSlideDeck/ batch generation). - Y4–Y6 — 15 Studio catalogue workflows under
examples/generators/(illustrated_slides,tech_slides_video,slides,video_overview,report,infographic,mindmap,circuit_diagram,audio_overview,flashcards,quiz,repo_architecture,backend_project_planner,paper_reader,video_summarizer),examples/applications/content-studio/, persistentnotebooks/sources/source_chunks/notebook_artifactswith unified citations andPOST /v1/notebooks/{id}/generate, plus inline workflowagentHITLask_userinterrupt/resume in gateway and Studio UI. - Y7–Y10 — cost-budget approvals,
ToolEffect::Publish, gateway reach ceilings, input precedence fixes, streaming thinking & reply images, Anthropic adaptive thinking + native structured output, and live multimodal provider verification.
Tracks that this product does not need yet (distributed workers, remaining
multi-replica gaps, coding-session Harness providers, seccomp refinement, MCP
follow-ups) are listed in RoadMap.md § Deferred, each with the condition
that reopens it.
The previously closed V (2026-08-05 production-readiness remediation)
segment came from findings by six independent parallel sub-agents that each
read one architecture layer, cross-checked by the orchestrator running
cargo test/clippy/fmt/tree locally:
- V0 — blocking correctness/security bugs (panics, unsafe defaults, terminal-state bugs, path traversal, an auth gap).
- V1 — core execution/agent robustness (DAG terminal-state + resume semantics, cancellation races, reliability-stack wiring, LLM retry, memory runtime integration, log normalization).
- V2 — structured output, token-level streaming, a generic HITL
interrupt/resume protocol (
ask_user), and agent-loop checkpointing. - V3 — fail-closed defaults on non-loopback binds, execution-side
sandbox enforcement for DAG
shellnodes, two SSRF gaps closed inHttpTool, per-tenant run admission control, supply-chain tightening (Ed25519 marketplace/plugin signatures, bounded MCP stdio reads), and anexprparser recursion-depth limit. - V4 —
AGENTS.mdregenerated from current sources,yanshi-cli’s 2700-linemain.rssplit into per-domaincommands/*/cli.rsmodules,cargo-auditwired into CI (11 of 13 pre-existing CVEs resolved via targeted dependency bumps), and completeness follow-ups: aninput_mappingwhitespace-parsing bug, a Geminibase_urlbug, DB-free~/.yanshi/runsretention, and (as a follow-up once the rest of V4 closed)chunk_strategy = "semantic"now reachable from a skill’s[[knowledge]]manifest entries, backed byyanshi-rag’s embedding-basedSemanticChunker.
The ongoing documentation-convergence convention:
- keep this file as the current status entrypoint;
- keep
RoadMap.mdfocused on future direction; - keep
TODOs.mdfocused on short-term execution; - leave historical evaluations marked as historical references.
Historical References
PROJECT_EVALUATION_2026-11-04.md: most recent evaluation — closure of the 10-07 findings (36 fixed / 13 partial / 17 open), new findings, then re-weighted for the content-production goal (A− library / B− platform); drove the Y track.PROJECT_EVALUATION_2026-10-07.md: evaluation that drove the X segment (gateway policy, cross-replica fixes).PROJECT_EVALUATION_2026-08-09.md: evaluation that drove the W segment; the baseline ofTRIAL_GUIDE.md.PROJECT_EVALUATION_2026-08-05.md: 6 parallel layer sub-agents + orchestrator cross-check, B+/A- composite, production-readiness C+; drove the V segment above.PROJECT_EVALUATION_2026-05-19.md: prior module-by-module evaluation (A overall, v1.0.0-rc.1 candidate).PROJECT_EVALUATION_2026-05-14.md: prior evaluation that informed the P6/P7/P-H/M segment additions.PROJECT_EVALUATION_2026-05-01.md: historical module-by-module evaluation that informed the P0-P4 task queue.RoadMap.md: roadmap and future direction.TODOs.md(maintainer-local, gitignored): active execution queue and task completion record.