lgd-theory
LGD — a registered · evidenced · gated lifecycle operational model (Registry / Evidence / Gate) for AI and autonomous systems:有籍 · 有证 · 有门禁。Theory SSOT. Concept DOI 10.5281/zenodo.22456647. 理论文本保留所有权利,注明出处可引用。
- 7
- HTML
- 5 h ago
LGD — a registered · evidenced · gated lifecycle operational model (Registry / Evidence / Gate) for AI and autonomous systems:有籍 · 有证 · 有门禁。Theory SSOT. Concept DOI 10.5281/zenodo.22456647. 理论文本保留所有权利,注明出处可引用。
XLGD — X distinction mark for LGD (Lifecycle Governance Doctrine). 凡自治之物 · 全程治理论: registry, evidence, gates. Maintained by SynomosAI Governance Line.
Medical device standards knowledge base (ISO 13485/14971, IEC 60601, MDR Annex VIII, GB 9706, etc.). Read-only MCP runtime for regulatory standards query across 27 knowledge hubs. Zero egress, zero cr
Run local LLMs & build a knowledge base on an old computer or low-spec PC - no GPU, no API key, offline, privacy-first. | SynomosAI skill - medxpert-llm-library - (c) 2026 SynomosAI - MIT
Candidate definitions for harmonization: silent-failure modes & lifecycle governance terms (CC BY 4.0). Pointed artifacts keep their own licenses.
Terminance (退役治理 / Retirement Governance) — LGD Theoretic Term. Part of the LingGongDao (凡自治之物) theory stack: what happens to an autonomous thing when it must stop. Theory-layer zero-collision coined term.
The Xinghua Cognitive Governance System (XCGS) - A Manifesto: AI Cognitive System, AI Identity Non-copyability, Theory Positioning
MedXpert 全球医械法规连接器 (reg-connector) — read-only MCP server, NMPA/FDA/MDR/PMDA knowledge hubs (27 pivots), 4 tools, zero egress. Install-by-consent (AI may request, human approves).
Medical device compliance tools. Self-assessment checklists for MDR / IVDR / SaMD / cybersecurity / post-market surveillance. Read-only, zero egress, audit-traceable.
Medical device testing coordination. Read-only tool for querying accredited testing labs, test method standards (IEC/ISO/GB), sample requirements and testing timelines across NMPA/FDA/MDR.
Medical device regulatory document translator. Read-only MCP tools for terminology lookup across NMPA/FDA/MDR/Japan PMDA — NOT a general translator. Audit-traceable, conservative translation stance.
One command, zero dependencies: grade any AI skill (OpenSSF Scorecard for agents). 一条命令零依赖,给任意 AI 技能打分。
Generate a single source of truth for your skill collection: scan skills dir into JSON + Markdown registry. 生成技能集合的单一真源:扫描技能目录 → JSON + Markdown 注册表。
Pre-release health check for AI skills: frontmatter completeness, placeholders, LGD markers, safety boundary. 技能发布前体检:frontmatter 完整性、占位词、三律标记、安全边界。
Scan AI outputs for silent failures: placeholders, empty stubs, hedging, no-progress repeats, hallucination markers. 扫描 AI 输出的静默失败:占位、空壳、含糊、原地重复、幻觉标记。
Detect prompt injection, hardcoded secrets, destructive commands and exfiltration patterns in any text or file. 检测提示注入、硬编码密钥、破坏性命令与数据外泄模式。
Generate llms.txt (LLMs.txt v2 spec) and AI-discoverability packages for any project, so AI agents can find and read your assets. 为任意项目生成 llms.txt(LLMs.txt v2 规范)与 AI 可发现性包,让 AI 代理能找到并读懂你的资产。
Generate ORCID-style fresh badges (20px, color-coded, thin type) for any project. 为任意项目生成 ORCID 清新风格徽章(20px、四色、细字)。
Medical device registration operational toolkit. Read-only MCP tools for querying NMPA/FDA/MDR submission paths, fee schedules, document checklists, and procedural timelines. Audit-traceable.
Medical device registration assistant (read-only MCP). Operates as a tool agent — does NOT provide medical advice or clinical judgment. For documentation query, regulatory path comparison, submission
LGD — Lifecycle Governance Doctrine. Public entry & contributor system: RUN IT / BREAK IT / BUILD IT. Theory SSOT: zhaoxinghua09-cell/lgd-theory · protocol: zhaoxinghua09-cell/uibc-core. 理论文本保留所有权利;代码 Apache-2.0。
ai agent — 指纹锁定交付 · © 2026 SynomosAI · MIT — 为任意交付物(HTML 白皮书/报告/文档)附加不可抵赖的内容指纹与出处锚定
Local-first, offline credential vault: consolidate passwords, API keys and tokens into one zero-knowledge, recoverable, tiered-authorization store with tamper-evident audit | SynomosAI skill - credential-vault-design - (c) 2026 SynomosAI - MIT
Guard AI context usage with a fail-closed token budget — stop burning tokens. 用 fail-closed 的 token 预算守护 AI 上下文,防止积分烧光。
ai agent — SynomosAI skill - authz-code-design - (c) 2026 SynomosAI - MIT
LGD-MedAI demo: evidence-gated lifecycle governance for a (fictional) medical AI agent. One command, zero dependencies.
Nomos - governance and rule-oriented work of the SynomosAI line; turning governance principles into operational instruments.
Runtigil (运行时门禁治理 / Runtime Gatekeeping) — LGD Theoretic Term. Gatekeeping enforced while an autonomous thing is running, not only at design time. Theory-layer zero-collision coined term.
Mnemoship (可携带记忆律 / Portable Memory Law) — LGD Theoretic Term. Portable memory for autonomous things: memory + identity that can move between runtimes without loss. Theory-layer zero-collision coined term.
Culpachain (多智能体问责治理 / Multi-Agent Accountability) — LGD Theoretic Term. Accountability chain across multiple autonomous agents: who answers when a chain of agents acts. Theory-layer zero-collision coined term.
Assurability (可保性治理 / Governance-Conditioned Insurability) — LGD Theoretic Term. Whether an autonomous thing is insurable is a governance outcome, not an actuarial accident. Theory-layer zero-collision coined term.
Assayance — the instrument is what gets assayed. Five criteria (T0–T5) for deciding whether a check is a check. The meta-method behind the LGD verification stack.