메뉴
Run shell commands and analyze output with validated summaries. Use for build logs, test output, or any command with substantial output. Protects context by returning concise summaries with grep-validated citations.
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
Use this when you need a complete interactive report as a single HTML file with zero install dependencies — covers charts, tables, diagrams, math, KPIs, and prose out of the box.
Use this as the default toolset for coding sessions when you want faster navigation, search, file inspection, and git workflow execution with lower command overhead: tools: - fzf - ripgrep - bat - delta - lazygit - starship - zoxide - eza - atuin - yazi
Out-of-band codebase exploration using a cheap/fast model. Keeps your context clean while a cheap model runs 30-80 commands in parallel. Tunable precision and output size - works for broad orientation or precise targeted questions.
Deterministically fixes broken OpenAI Deep Research markdown citations without using an LLM: creates a .bkp backup, rewrites citation markers, rebuilds references, and runs strict regex validation.
Explore a codebase to gather context for coding tasks. Preserves context window and offloads work to cheaper models.
Dynamically generate and execute a specialized agent for any task. Use when the user has an arbitrary task that would benefit from a custom-generated agent.
| name | shell-analyzer |
| description | Run shell commands and analyze output with validated summaries. Use for build logs, test output, or any command with substantial output. Protects context by returning concise summaries with grep-validated citations. |
Run immediately with the shell command to execute:
$HOME/.agents/skills/shell-analyzer/run.sh "<command>"
Use --style=<style> to control output verbosity:
| Style | Description | Use Case |
|---|---|---|
compact | Brief summary + key findings (default) | General use |
minimal | Single line status | Quick checks |
detailed | Full markdown sections | Debug/investigation |
errors-only | Only output if problems found | CI/build monitoring |
# Default compact style
run.sh "npm run build"
# Minimal - just pass/fail
run.sh --style=minimal "cargo test"
# Detailed - full analysis
run.sh --style=detailed "pytest -v"
# Errors only - silent on success
run.sh --style=errors-only "make check"