← 返回 Skills 市场
Refines ambiguous or high-risk user requests before execution. Trigger when the request is underspecified, likely to benefit from clearer constraints or verification, or when the user explicitly asks to refine, improve, optimize, refactor, or compare approaches. Skip clear single-step instructions and already-well-scoped tasks.
作者
TaiChangXieBuWan
· GitHub ↗
· v1.0.3
· MIT-0
116
总下载
0
收藏
0
当前安装
3
版本数
在 OpenClaw 中安装
/install self-improving-prompt
功能描述
Refines ambiguous or high-risk user requests before execution. Trigger when the request is underspecified, likely to benefit from clearer constraints or veri...
安全使用建议
This skill appears coherent and low-risk, but double-check two operational details before enabling it broadly: (1) confirm how 'session context' is scoped and whether that context can include sensitive or confidential user data — ensure the platform prevents inadvertent exposure to other subsystems or logs; (2) verify that the 'minimal preference event' sent to self-improving-session is truly a small label (choose_refined / choose_original, etc.) and never contains the full prompt or other identifying data. Also confirm the platform tooling used for confirmations (AskUserQuestion fallback) behaves as expected for your privacy/compliance requirements.
功能分析
Type: OpenClaw Skill
Name: self-improving-prompt
Version: 1.0.3
The 'self-improving-prompt' skill is a purely instructional bundle (Markdown-only) designed to help an AI agent refine ambiguous or high-risk user requests. It contains no executable code, scripts, or network requirements, and includes clear logic (SKILL.md, decision-matrix.md) to avoid unnecessary friction or loops. There are no indicators of data exfiltration, malicious execution, or harmful prompt-injection instructions.
能力评估
Purpose & Capability
The name/description match the runtime instructions and reference materials. It only needs to read the user's request and session context to decide whether to refine — which is coherent for a prompt-refinement skill. No unrelated binaries, env vars, or install steps are requested.
Instruction Scope
Instructions require reading the original request and current session context (expected for this task). The skill also asks to record minimal preference events for later summarization by a separate 'self-improving-session' skill; this is reasonable but worth auditing to ensure no full prompts or sensitive content are stored or leaked.
Install Mechanism
Instruction-only skill with no install spec, no downloads, and no bundled code — lowest-risk installation profile.
Credentials
No environment variables, credentials, or config paths are requested. Requested access (session context and the immediate user request) is proportional to the skill's purpose.
Persistence & Privilege
always:false and default model invocation are appropriate. The skill does ask to emit small preference events for later summarization, but it explicitly warns not to store full refined prompts; this is consistent with limited persistence and limited privilege.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install self-improving-prompt - 安装完成后，直接呼叫该 Skill 的名称或使用
/self-improving-prompt触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.3
**Summary:** Improves clarity, selectivity, and decision structure for when and how prompt refinement is applied.
- Refines ambiguous or high-risk user prompts only when material value (such as clearer scope or acceptance criteria) would result; skips clear or atomic instructions
- Introduces a decision matrix (`references/decision-matrix.md`) to guide between direct execution, silent refinement, or compare-first modes
- Adds concrete examples of cases that do NOT warrant refinement (`references/non-examples.md`)
- Tightens the definition of "substantial value"—now requires refinement to clarify at least two core aspects (goal, scope, output, etc.) to prompt user comparison
- Clarifies integration with self-improving-session, with stricter event logging and privacy: never store full prompt text, only workflow choice signals
- Streamlines output modes and preference event rules to minimize user interruption
v1.0.2
**Skill self-improving-prompt v1.0.2 Changelog**
- Removed README.md and scripts/self-improving-prompt-hook.sh files.
- Updated SKILL.md with new Setup instructions for integrating with `~/.claude/settings.json`.
- No changes to skill logic or core functionality.
v1.0.1
**Prompt-refiner renamed and improved for clarity and modularity:**
- Renamed skill from `prompt-refiner` to `self-improving-prompt` for consistency.
- Improved initial trigger rules and continue modes; updated terminology for greater clarity.
- Output templates and workflow clarified; modular structure for refined prompts specified.
- Integration points renamed (`self-improving-session`) and preference signaling outlined.
- Added detailed README and reference files for usage and guidance.
元数据
常见问题
Refines ambiguous or high-risk user requests before execution. Trigger when the request is underspecified, likely to benefit from clearer constraints or verification, or when the user explicitly asks to refine, improve, optimize, refactor, or compare approaches. Skip clear single-step instructions and already-well-scoped tasks. 是什么？
Refines ambiguous or high-risk user requests before execution. Trigger when the request is underspecified, likely to benefit from clearer constraints or veri... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 116 次。
如何安装 Refines ambiguous or high-risk user requests before execution. Trigger when the request is underspecified, likely to benefit from clearer constraints or verification, or when the user explicitly asks to refine, improve, optimize, refactor, or compare approaches. Skip clear single-step instructions and already-well-scoped tasks.？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install self-improving-prompt」即可一键安装，无需额外配置。
Refines ambiguous or high-risk user requests before execution. Trigger when the request is underspecified, likely to benefit from clearer constraints or verification, or when the user explicitly asks to refine, improve, optimize, refactor, or compare approaches. Skip clear single-step instructions and already-well-scoped tasks. 是免费的吗？
是的，Refines ambiguous or high-risk user requests before execution. Trigger when the request is underspecified, likely to benefit from clearer constraints or verification, or when the user explicitly asks to refine, improve, optimize, refactor, or compare approaches. Skip clear single-step instructions and already-well-scoped tasks. 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Refines ambiguous or high-risk user requests before execution. Trigger when the request is underspecified, likely to benefit from clearer constraints or verification, or when the user explicitly asks to refine, improve, optimize, refactor, or compare approaches. Skip clear single-step instructions and already-well-scoped tasks. 支持哪些平台？
Refines ambiguous or high-risk user requests before execution. Trigger when the request is underspecified, likely to benefit from clearer constraints or verification, or when the user explicitly asks to refine, improve, optimize, refactor, or compare approaches. Skip clear single-step instructions and already-well-scoped tasks. 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Refines ambiguous or high-risk user requests before execution. Trigger when the request is underspecified, likely to benefit from clearer constraints or verification, or when the user explicitly asks to refine, improve, optimize, refactor, or compare approaches. Skip clear single-step instructions and already-well-scoped tasks.？
由 TaiChangXieBuWan（@lq434239）开发并维护，当前版本 v1.0.3。
推荐 Skills