Lingtai-AI/lingtai-kernel

system-settings-inventory-reference

Exact read-only System catch-all settings inventory: ownership, effective sources, defaults, accepted values, invalid behavior, redaction, timing, authorized change procedures, and explicit non-settings.

system-manual

Short router for runtime, lifecycle, identity, Agent/Project relocation, refresh transactions, presets, settings, update/mismatch diagnosis, LLM adapters, and operating procedures.

context-manual

Use for an unfamiliar or consequential context operation, session journaling, handoff, or context-loss recovery; routine calls can follow the schema.

daemon-manual

Use for unfamiliar daemon delegation, backend-specific work, stalled or failed runs, completion checks, or consent-gated cleanup. Routine calls can start from the advertised schema.

workflow-manual

Ordered daemon batches from a snapshotted JSON workflow file.

daemon-backend-claude-p

Nested daemon-cli-backends reference for the claude-p (alias claude-code) daemon backend's flag surface. Read this only when a daemon task needs Claude Code-specific CLI flags (model selection, fallback model, tool restrictions): it routes you to the…

daemon-cli-backends

Nested daemon-manual reference for backend choice, native versus prompt-only MCP, CLI options, list/ask behavior, and per-backend live-help routes.

recipe-export-flow

Nested lingtai-recipe reference for exporting a standalone recipe bundle from a live network or methodology: scope disambiguation, human-in-the-loop metadata collection, bundle authoring, validation, git initialization, and handoff.

recipe-format-reference

Nested lingtai-recipe reference for authoring and validating recipe bundles: bundle structure, .recipe/recipe.json schema, behavioral layers, library sibling mechanics, locale fallback, validator checks, hand-authored recipes, testing, and publishing.

lingtai-recipe

Menu manual (not a tool) for recipes — the named payload that shapes an orchestrator's greeting, ongoing behaviour, and shipped library; every LingTai project uses one. Routes to `reference/recipe-format/SKILL.md` (authoring/customising) and…

telegram-mcp-manual

Progressive-disclosure usage manual for the Telegram MCP tool. The resident schema carries safe first-use guidance; call `manual` for the action map, inbound-first/reply routing, channel/media/rendering rules, settings, Task Card projection, and error…

telegram-task-card-projection-notice

Shipped retained-legacy/projection notice for Telegram Task Card files. The public `task_card` tool is intrinsic and documented at src/lingtai/tools/task_card/manual/SKILL.md; Telegram only projects the intrinsic taskcard/status + taskcard/taskcard.md…

whatsapp-mcp-manual

Progressive-disclosure usage manual for the personal-account WhatsApp MCP tool. Read this when you need detail beyond the concise action description: the single-session whatsapp-web.js bridge and QR pairing, strict send/reply/react semantics, opaque message…

feishu-mcp-manual

Concise progressive-disclosure usage manual for the Feishu (Lark) MCP. Ordinary schema-sufficient calls can start here; unfamiliar setup, cards, callbacks, compound replies, partial delivery, and recovery route to one focused reference. It covers the strict…

psyche-manual

Compact router for Psyche's private-memory and procedure domains (pad, knowledge, skills), the shared file→rebuild model, redacted settings, and the separate fixed-instructions disclosure and retirement guidance.

shell-manual

**Read before running a long-lived agent/coding CLI as a shell subprocess**, or before setting up cron/launchd/systemd timers or scheduled reminders. Routes shell-side async+poll supervision, host-scheduler setup, LingTai wake-by-mailbox-drop, one-shot…

summarize-manual

Operational guide for recording tool-result summaries, applying them with a full rebuild, recovering raw evidence, and choosing summarize versus molt.

lingtai-manual

Retirement reference for legacy character memory; use unified Pad and Knowledge.

pad-manual

Read before editing the living Pad, pinning references, or preparing Pad state for rebuild/molt.

procedures-manual

Nested system-manual reference: the expanded form of the resident procedures prompt — progressive disclosure, responsiveness and side-effect authorization, the daemon workflow methodology, depositing work, idle/lifecycle, skill routing, deliverables, artifact…

substrate-manual

Nested system-manual reference: the expanded form of the resident substrate prompt — body/extensions, lifecycle states, the `system` tool in practice, communication and memory layers, collaboration topology, MCP/addon ownership, and (§11) the canonical…

avatar-manual

Use for an unfamiliar or consequential Avatar spawn, shallow/deep choice, detached lifecycle, recovery, or footprint review. Routine schema-sufficient spawns can use the tool directly.

skills-manual

Read before authoring, installing, or sharing a skill in `.library/custom/<name>/`, adding configured skill roots, or diagnosing a catalog entry; this teaches the Skills system, not bundled skill procedures.

plugin-manual

Read-only Agent Plugins v1.0.0 catalog: registered versus discovered, authorized registration refresh, containment, and skipped-component diagnosis. Use for unfamiliar or consequential plugin work; routine schema-sufficient calls can go straight to `info`.

cloud-mail-mcp-manual

Progressive-disclosure usage manual for Cloud Mail REST email actions. Read it when action filters/ids, setup/authentication, polling/settings, or side-effect detail is needed. Returned by action='manual'; nested references are packaged but not embedded.…

imap-mcp-manual

Progressive-disclosure usage manual for the IMAP/SMTP email MCP. Read this router for unfamiliar or consequential workflows and deeper detail on real-mail side effects, external-reply policy, account selection, compound email IDs, attachments, mailbox…

wechat-mcp-manual

Progressive-disclosure usage manual for WeChat/iLink: first safe read, exact IDs, external send/reply, media partials, and standalone poller/setup hazards. Use the schema for routine calls; load this manual for media details, unfamiliar results, setup, or…

web-manual

First-call search→browse route for the single web capability; exact settings, complete-output, provider, and public-URL hazards are retained here or routed to one focused reference.

lingtai-doctor

Read-only health diagnostics when an agent looks dead but the evidence is mixed, or a migration may have left stale MCP/addon paths; run before deciding to mail, refresh, CPR, or edit persistent configuration. Includes a bundled read-only `doctor.py`.

mcp-manual

Read-only MCP status router: distinguish catalog, registered, and active servers; inspect effective settings safely; and route authorized setup or recovery to the exact provider and runtime references. Routine info/settings/manual calls need no ritual manual…

lingtai-kernel-anatomy

Enter the Python kernel Anatomy/Contract graph before source navigation or maintaining architecture maps; explains how the ANATOMY.md and CONTRACT.md graphs pair, and what to do when code and navigation disagree.

psyche-settings-reference

Deep reference for Psyche's redacted settings SHOW, strict owner-document grammar, precedence, timing, rollback, and six stable anchors.

llm-adapters

Nested system-manual reference for LingTai's four built-in LLM provider families (openai, anthropic, codex, claude-code): what each adapter is, how it is configured and dispatched, its standard parameters (wire_api, thinking, service_tier), how other vendors…

migration-agent-name-move

Narrow contract for the self-contained POSIX same-parent Agent workdir/address rename script and its transactional runtime/MCP relocation.

migration-project-move

Narrow POSIX contract for the bundled initiating-Agent whole-Project same-volume self-move helper.

migration-guide

Conclusion-first router for authorizing, selecting, recovering, and proving a mechanical Agent workdir/address or whole-Project path move.

runtime-update-checks

Nested system-manual reference for update, source, install, nudge, and runtime-mismatch diagnosis. Owns interpreter/provenance evidence and the source/venv cutover handoff; routes refresh sequencing to refresh-precheck.

tool-plugin-settings-reference

Developer reference for opting one tool family into bounded, read-only settings discovery without changing other families.

telegram-rate-limits

Current official Telegram Bot API flood-control guidance: published quotas, `retry_after` semantics, documented unknowns, and safe client policy. Read before changing Telegram send cadence, programmable Task Card cadence, or 429 recovery behavior.

daemon-backend-qwen-code

Nested daemon-cli-backends reference for the Qwen Code (`qwen-code` / `qwen`) daemon backend's flag surface. Read this only when a daemon task needs Qwen-specific CLI flags (model selection, provider tunables): it routes you to the installed CLI's live help…