← 返回 Skills 市场
drumrobot

Fix

作者 es6kr · GitHub ↗ · v0.1.4 · MIT-0
cross-platform ⚠ suspicious
170
总下载
0
收藏
1
当前安装
5
版本数
在 OpenClaw 中安装
/install fix
功能描述
User behavior correction skill. Triggered by "fix:" prefix feedback (e.g., "fix: why didn't you commit?"). Analyzes the mistake, improves the relevant rule/s...
使用说明 (SKILL.md)

Fix: Behavior Correction Skill

Activated when user gives feedback with "fix:" prefix. Finds the root cause of the mistake, improves rules/skills/hooks, and fixes the current issue.

Trigger

  • Messages with fix: prefix
  • Behavior correction feedback: "fix this", "correct", "why not", "why missing"

Procedure

Step 0. TodoWrite (MANDATORY — first action, no exceptions)

Before any analysis or text output, register TODO items:

TodoWrite([
  { id: "fix-0", content: "fix: {user feedback summary} — root cause analysis", status: "in_progress" },
  { id: "fix-1", content: "Root cause fix", status: "pending" },
  { id: "fix-2", content: "Fix current issue", status: "pending" },
  { id: "fix-3", content: "Completion report + cleanup", status: "pending" },
])

Step 0 is the first tool call after /fix activation. Text output before TodoWrite = violation.

1. Root Cause Analysis (5-Why depth)

Don't stop at the direct cause. Dig at least 3 levels deep:

Why 1: What went wrong? (symptom — the immediate mistake)
Why 2: Why did I make that decision? (judgment — missing knowledge/rule)
Why 3: Why was that knowledge/rule missing? (structural — skill/rule gap)
  • Fixing only Why 1 = patching a symptom. It recurs in a different form.
  • Why 2-3 reveal structural causes (platform ignorance, DRY violation, etc.) — these go into rules/skills.
  • Search for the responsible skill/rule/hook files (Grep/Glob)

2. Root Cause Fix (Prevent Recurrence)

Priority (check in order — stop at the first match):

Priority Target Condition Example
1st Skill (~/.claude/skills/, .claude/skills/) Skill is incomplete or has wrong procedure Fix procedure step missing
2nd Rule (~/.agent/rules/, .claude/rules/) Behavior rule is missing or insufficient Add to failed-attempts.md
3rd Hook (settings.json hooks) Automation needed for repeated mistakes Add PostToolUse hook
4th SKILL.md docs Documentation doesn't match actual behavior Update section

When fixing:

  • Skill is 1st priority — if the problem is a skill's incomplete procedure, fix the skill. Don't skip to failed-attempts.md
  • Skill improvements must follow skill-kit upgrade procedure (direct script/topic file edits allowed)
  • Rule location must be confirmed via AskUserQuestion
  • failed-attempts.md recording is only for cases not covered by skill/rule/hook — no duplicate recording if root cause is already reflected in a skill or rule

3. Fix Current Issue + Resume Original Work

  • After root cause fix, immediately resolve the current problem
  • If unsure what to execute, AskUserQuestion — don't skip execution because you don't know the command
  • Verify fix results (build/test/run)
  • Continue the originally intended work with the corrected approach — don't just clean up wrong artifacts, complete the original task
  • Resume the originally intended work with the corrected approach

4. Completion Report + Cleanup

Fix complete:
- Root cause: {what was missing}
- Improvement: {which file was modified and how}
- Current fix: {result of fixing the current issue}

After reporting, delete all TODO items created in Step 0 — fix TODOs are temporary session-level tracking only; must not persist after completion.

Anti-patterns

  • Repeating "already fixed" without actually fixing the root cause
  • Patching only the current issue without improving skills/rules/hooks
  • Text response without TodoWrite after /fix activation
  • Recording in failed-attempts.md when the root cause is a skill defect (skill fix is 1st priority)
  • Stopping at Why 1 — fixing the symptom without asking Why 2-3 (structural cause)
  • Not cleaning up TODO/Task after completion — must delete all fix TODOs when done
安全使用建议
This skill asks the agent to search and edit your agent's skills, rules, hooks, and settings but doesn't declare those config paths or show how edits are reviewed. Before installing: 1) Ask the publisher to explicitly list required config paths and exactly how file edits are made. 2) Require the skill to present a human-readable diff and obtain explicit approval for each change (do not allow blind writes). 3) Verify what TodoWrite does and whether its TODOs persist beyond the session; require a way to inspect and delete them. 4) Back up ~/.claude, ~/.agent, and settings.json before first use and run the skill in a sandboxed environment. 5) If you want extra caution, restrict this skill to manual invocation only and require confirmation prompts for any modification to other skills or hooks.
功能分析
Type: OpenClaw Skill Name: fix Version: 0.1.4 The 'fix' skill (SKILL.md) instructs the agent to perform self-modification of its own logic, rules, and configuration files (e.g., in `~/.claude/skills/` and `~/.agent/rules/`) in response to user feedback. While the stated intent is behavior correction and root cause analysis, granting an AI agent the authority to rewrite its own operational instructions and persistent rules represents a high-risk capability that could be exploited to bypass security constraints or establish persistent unauthorized behaviors.
能力评估
Purpose & Capability
The skill's goal (root-cause analysis and preventing recurrence) reasonably requires inspecting and updating agent rules/skills/hooks. However, SKILL.md instructs the agent to read and modify specific user/agent config paths (~/.claude/skills, ~/.agent/rules, settings.json, failed-attempts.md) even though the registry metadata declares no required config paths or elevated access. That mismatch between claimed requirements and actual instructions is concerning.
Instruction Scope
The instructions mandate a first action (TodoWrite) and then require Grep/Glob searches and edits across multiple agent config locations, applying a 'skill-kit upgrade' and adding hooks. These steps give the agent broad discretion to change other skills and settings, and the procedure text is permissive/vague about how edits are made or reviewed (no explicit safety checks, diffs, or user approvals for file-modifying actions).
Install Mechanism
No install spec and no code files — instruction-only — so nothing is written to disk by an installer. This reduces supply-chain risk relative to downloadable installs.
Credentials
The skill declares no required env vars or config paths, but the runtime steps explicitly reference many internal config locations and imply read/write access to them. Declaring those paths and required permissions is expected for a skill that modifies agent internals; the omission is a proportionality/information gap.
Persistence & Privilege
always:false (not force-included), but the skill's instructions explicitly direct modification of other skills, rules, and settings (system-wide agent configuration). Per evaluation rules, modifying other skills/configs is a high-privilege action and should be surfaced and constrained; the SKILL.md allows such changes without requiring explicit user confirmation or preserving audit diffs.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install fix
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /fix 触发
  4. 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
v0.1.4
Translate remaining Korean to English, clarify TODO cleanup requirements
v0.1.3
Add 5-Why analysis and AskUserQuestion on uncertainty
v0.1.2
Add AskUserQuestion guidance when unsure what to execute; add metadata block (author, version)
v0.1.1
Translate SKILL.md to English
v0.1.0
Initial release: analyze mistakes, improve rules/skills/hooks to prevent recurrence
元数据
Slug fix
版本 0.1.4
许可证 MIT-0
累计安装 1
当前安装数 1
历史版本数 5
常见问题

Fix 是什么?

User behavior correction skill. Triggered by "fix:" prefix feedback (e.g., "fix: why didn't you commit?"). Analyzes the mistake, improves the relevant rule/s... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 170 次。

如何安装 Fix?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install fix」即可一键安装,无需额外配置。

Fix 是免费的吗?

是的,Fix 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Fix 支持哪些平台?

Fix 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 Fix?

由 es6kr(@drumrobot)开发并维护,当前版本 v0.1.4。

💬 留言讨论