Diagnose Openclaw Model Routing And Fallback Usage
v1.0.0用于分析 OpenClaw 中“某个 provider/模型为什么看起来没被用上”、主模型与 fallback 实际是否生效、以及 ClaudeCodeCLI 协作任务为何中断。遇到这些情况就应触发：用户提到“没请求打到 local-router”“明明注册了模型却没流量”“Agent 到底在用哪个模型”“fal...
⭐ 0· 73·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description match the runtime instructions: the SKILL.md focuses on correlating OpenClaw agent configuration, provider/model registration, fallback chains, and logs plus diagnosing ClaudeCodeCLI SIGTERM/143 failures. All required actions (reading configs/logs, checking fallback decision fields, inspecting CLI exit codes, suggesting safer CLI invocation patterns) are appropriate for this purpose.
Instruction Scope
The instructions explicitly tell the agent to read OpenClaw configuration and log files (examples include config YAMLs and paths like /var/log/openclaw/server.log), examine fallback decision fields, and run or recommend running local CLI commands (claude -p). These actions are sensitive but directly relevant to diagnosing routing/fallback and CLI termination; the skill does not instruct collection or transmission of unrelated system secrets or exfiltration to external endpoints. Users should be aware the skill assumes access to local files and system logs.
Install Mechanism
Instruction-only skill with no install spec and no code files. No downloads, no package installs, and nothing is written to disk by the skill itself beyond advising best-practice commands (e.g., redirecting CLI output to /tmp). This is the lowest-risk install mechanism.
Credentials
The skill declares no required environment variables, no credentials, and no config paths in the metadata. The SKILL.md does reference local endpoints (e.g., http://localhost:8317/v1) and local file paths for logs/configs — all are directly relevant to the diagnostic purpose. It does not request unrelated API keys or secrets.
Persistence & Privilege
The skill is not always-enabled (always: false) and is user-invocable. It does not request persistent system-level privileges or attempt to modify other skills' configurations. It can be invoked autonomously by default platform behavior, but there are no other elevated persistence demands in the skill bundle.
Assessment
This skill is coherent and appears to do what it claims: correlate OpenClaw configs and logs and diagnose ClaudeCodeCLI SIGTERM/143 issues. Before installing, consider: 1) the skill's runtime instructions assume the agent can read local config and log files (e.g., /var/log/openclaw/server.log, config YAMLs, systemd/journalctl), so only install if you trust the agent to access those files; 2) it recommends running local CLI commands (claude -p) and writing output to /tmp — ensure those binaries are present and you consent to such operations; 3) the skill does not request external credentials or network exfiltration, but it will surface potentially sensitive log contents when run; 4) if you prefer to avoid automated reads of system logs, restrict the skill to manual/user-invoked use or review its text before each run; and 5) keep normal operational safeguards (run in a controlled environment, audit outputs, and do not grant it access to secrets you wouldn't want examined).Like a lobster shell, security has layers — review code before you run it.
latestvk9734tbms7348waymtxt480v1x83cv7a
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🧭 Clawdis