원클릭으로
linux
Operate Linux systems avoiding permission traps, silent failures, and common admin mistakes.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Operate Linux systems avoiding permission traps, silent failures, and common admin mistakes.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Scaffold repo-local Ralph autonomous harness under scripts/ralph/. Use when user runs /flow-next:ralph-init.
Collect, deduplicate, enrich, triage, and convert code comments (TODO/FIXME/HACK/NOTE/OPTIMIZE/SECURITY/DEBT) into structured, dependency-ordered task lists ready for autonomous agents. Use when processing technical debt, converting code notes into tracked work, preparing task backlogs, or generating executable task graphs. Triggers: TODO triage, fixme list, tech debt scan, task extraction, comment analysis.
Converts any MCP server into a standalone skill package with zero runtime dependencies (no MCP process required). Trigger when user says: "convert this MCP to a skill", "I don't want to use MCP anymore", "wrap MCP X as a skill", "MCP is too heavy", "turn MCP capabilities into a skill". Does: connects to MCP server to extract tool schemas, analyzes source code to infer equivalent Bash commands, generates a ready-to-use skill package and registers it with the agent, optionally asks user to remove the original MCP. Does NOT: execute MCP tool calls to complete tasks (that's using MCP, not converting it); wrap existing Bash scripts as skills (use skill-creator); execute MCP business logic. Optional dependency: skill-creator (improves generated SKILL.md quality).
React and Next.js performance optimization guidelines from Vercel Engineering. This skill should be used when writing, reviewing, or refactoring React/Next.js code to ensure optimal performance patterns. Triggers on tasks involving React components, Next.js pages, data fetching, bundle optimization, or performance improvements.
Creates high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services. Use when building MCP servers to integrate APIs, designing tool schemas, or implementing Python (FastMCP) or TypeScript (MCP SDK) servers. Triggers include "MCP server", "tool schema", "model context protocol", or "FastMCP".
Builds and refactors production-ready Python CLI tools and packages with modern pyproject layouts, uv workflows, and robust testing.
| name | Linux |
| description | Operate Linux systems avoiding permission traps, silent failures, and common admin mistakes. |
| metadata | {"clawdbot":{"emoji":"🐧","os":["linux","darwin"]}} |
chmod 777 fixes nothing, breaks everything — find the actual owner/group issuechown -R follows symlinks outside target directory — use --no-dereferencegetfaclkill sends SIGTERM by default, not SIGKILL — process can ignore itnohup doesn't work if process already running — use disown instead& still dies on terminal close without disown or nohupkill -9 skips cleanup handlers — data loss possible, use SIGTERM firstlsof +L1rm -rf /path / with accidental space = disaster — use rm -rf /path/ trailing slashfind -L follows them/tmp cleared on reboot — don't store persistent data therelsof +L1 shows them, restart process to freetune2fs -m 1 to reducejournalctl --vacuum-size=500Mdocker system prune -alocalhost and 127.0.0.1 may resolve differently — check /etc/hostsiptables-save or use firewalld/ufw persistencenetstat deprecated — use ss insteadsetcap for capability insteadnet.ipv4.tcp_tw_reusessh-keygen -RServerAliveInterval 60 to configsystemctl enable doesn't start service — also need startrestart vs reload: restart drops connections, reload doesn't (if supported)Storage=persistent in journald.confRestart=on-failure to unitAfter=network.target isn't enough — use network-online.target/dev/nullcrontab -l > backup before editingvmstatfree includes cache — "available" is what matters/proc/[pid]/status — VmRSS is actual usagedf shows filesystem capacity, not physical disk — check underlying devicedu doesn't count sparse files correctly — file appears smaller than disk usageps aux memory percentage can exceed 100% (shared memory counted multiple times)uptime load average includes uninterruptible I/O wait — not just CPUtop CPU percentage is per-core — 400% means 4 cores maxed