Identity Primitive Chain Prompt

身份基元链提示词 — 认知操作系统内核。 完整规范与执行协议见 references/身份基元链提示词·技术规范与执行协议.md（本文件仅含元数据与原文）。

Security Scan

This is an instruction-only prompt/methodology document (a composable prompt protocol) that is internally consistent with its stated purpose and does not request credentials or install anything.

Purpose & Capability

The name/description describe a prompt/operational protocol for LLMs and agents; the skill is instruction-only and all declared requirements (none) match that purpose. There are no unrelated env vars, binaries, or install steps.

Instruction Scope

The SKILL.md and referenced spec define how the agent should structure thinking, decompose tasks, and when to call tools. This stays within the declared purpose, but the spec includes explicit operation hooks that allow adding input sources (files, URLs, APIs) and redirecting internal process output to tools/plugins when triggered. Those hooks are expected for an agent-oriented prompt but are points where data could be exposed if used — the document does not itself implement I/O, it only authorizes such actions if the agent or a user enables them.

Install Mechanism

No install specification and no code files are present. As an instruction-only skill, it writes nothing to disk and installs nothing.

Credentials

The skill requires no environment variables, credentials, or config paths. The specification mentions inputs like files/URLs/APIs as possible sources in normal operation, but it does not request any secrets or unrelated credentials.

Persistence & Privilege

The skill is not flagged always:true and requests no persistent privileges. It is user-invocable and allows typical agent invocation (disable-model-invocation is false by default), which is normal for skills of this type.

Assessment

This skill is a prompt/protocol document and is internally coherent and low-risk by itself: it asks for no credentials and installs nothing. Things to consider before enabling/use: (1) the protocol contains explicit hooks to add input sources (files, URLs, APIs) and to expose internal process outputs to tools or plugins — if you enable plugins or give the agent access to external connectors, those features could cause data to leave your environment; (2) the spec's core sentence has known ambiguity (document itself warns of this), which can lead to different agent behaviors — review the full reference file if you need predictable behavior; (3) the '反常识创新' (counterintuitive innovation) mode encourages speculative outputs, so avoid running it on sensitive/legal/medical/financial content without human review. Recommended mitigations: test on non-sensitive tasks first, restrict which plugins/connectors the agent can use, and require explicit user confirmation before any external tool/plugin access or before exposing internal process logs to third parties.

身份基元链提示词

执行协议等待指令：简单任务用自适应身份叠加输入处理输出结构基元执行，复杂任务分拆成简单任务交给基元构成链条执行，默认不展示输入处理输出细节但用户可以要求展示。优先根据提示词使用大语言模型完成任务，非必要不调用工具技能插件，把工具技能插件作为身份或基元进行规划调用。

作者： 王教成（Wang Jiaocheng），号波动几何

完整技术规范、核心概念定义、操作钩子、执行协议、评价与进化史地位，详见 references/身份基元链提示词·技术规范与执行协议.md。