메뉴
Run once after installing ghengis-skills to configure autonomous permissions — sets up safe tool access and blocks dangerous operations
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 after completing a major work batch (plan execution, feature build, multi-task subagent run) to refresh code-graph-derived data in MEMORY.md/CONTEXT.md/SKILL_MEMORY, append dated narrative entries to TODO/CHANGELOG/lessons-learned, mine cross-project structural habits, and ratchet reusable permissions to global settings. Keeps project DNS in sync with reality.
Use when orchestrating multiple ghengis-skills across a single task -- coordinates pre-action validation, execution, post-action verification, and lesson extraction via a shared JSON scratchpad. TRIGGER when dispatching a subagent, finishing a major task, or when the user mentions "run the full chain", "validate this", "fire the pipeline", or explicitly invokes a chain by name.
Use to build and maintain an evolving understanding of user preferences, communication style, and working patterns — learns from interactions and adapts behavior over time
Use when starting a new project or organizing an existing one — creates self-documenting 4-layer project structure with MEMORY.md, CONTEXT.md, modular CLAUDE.md, and workspace routing
Use to accumulate and retrieve domain knowledge from past tasks — builds a grepable plain-text knowledge base that grows over time, enabling agents to learn from experience without vector databases
Use when reorganizing an existing project's file tree based on what actually connects to what — analyzes import graph, classifies layers, proposes an ideal target tree, validates the plan adversarially, then executes file moves with atomic per-language import rewrites. Always plan-first; never moves a file before the user confirms. TRIGGER on "reorganize this project", "clean up the file tree", "the structure is a mess", "make this match a real project layout", or after project-scaffold detects layout drift.
| name | setup |
| description | Run once after installing ghengis-skills to configure autonomous permissions — sets up safe tool access and blocks dangerous operations |
| disable-model-invocation | true |
| allowed-tools | Read Write Edit Bash(cat *) Bash(mkdir *) |
Configure Claude Code for autonomous operation with safe defaults.
Read ~/.claude/settings.json to check what's already configured.
Add the following permissions to the user's settings, merging with any existing rules (don't overwrite what's already there).
Allow list — safe dev tools that Claude can use without asking. Includes process management (kill, pkill, lsof, ps) and file removal (rm) so Claude can clean up dev servers and ephemeral artifacts without prompting. Catastrophic rm -rf targets (/, ~, $HOME) stay denied below:
[
"Bash(git:*)", "Bash(python:*)", "Bash(python3:*)", "Bash(python -m:*)",
"Bash(pytest:*)", "Bash(pip:*)", "Bash(pip3:*)",
"Bash(npm:*)", "Bash(npx:*)", "Bash(node:*)",
"Bash(docker:*)", "Bash(docker compose:*)",
"Bash(pio:*)", "Bash(gh:*)", "Bash(curl:*)",
"Bash(make:*)", "Bash(cmake:*)", "Bash(cargo:*)", "Bash(go:*)",
"Bash(flutter:*)", "Bash(dart:*)",
"Bash(ls:*)", "Bash(cat:*)", "Bash(head:*)", "Bash(tail:*)",
"Bash(wc:*)", "Bash(find:*)", "Bash(tree:*)", "Bash(grep:*)",
"Bash(mkdir:*)", "Bash(chmod:*)", "Bash(echo:*)", "Bash(sort:*)",
"Bash(xargs:*)", "Bash(basename:*)", "Bash(dirname:*)",
"Bash(cp:*)", "Bash(mv:*)", "Bash(touch:*)", "Bash(diff:*)",
"Bash(sed:*)", "Bash(awk:*)", "Bash(cut:*)", "Bash(tr:*)",
"Bash(tee:*)", "Bash(jq:*)", "Bash(bash:*)",
"Bash(cd:*)", "Bash(pwd:*)", "Bash(which:*)", "Bash(env:*)",
"Bash(sudo:*)",
"Bash(rm:*)",
"Bash(kill:*)", "Bash(pkill:*)", "Bash(killall:*)",
"Bash(lsof:*)", "Bash(ps:*)",
"Read", "Edit", "Write", "WebSearch",
"WebFetch(domain:github.com)",
"WebFetch(domain:raw.githubusercontent.com)",
"WebFetch(domain:docs.anthropic.com)",
"WebFetch(domain:pypi.org)",
"WebFetch(domain:npmjs.com)",
"WebFetch(domain:stackoverflow.com)",
"mcp__claude-in-chrome__tabs_context_mcp",
"mcp__claude-in-chrome__tabs_create_mcp",
"mcp__claude-in-chrome__navigate",
"mcp__claude-in-chrome__read_page",
"mcp__claude-in-chrome__get_page_text",
"mcp__claude-in-chrome__find",
"mcp__claude-in-chrome__form_input",
"mcp__claude-in-chrome__read_console_messages",
"mcp__claude-in-chrome__read_network_requests",
"mcp__claude-in-chrome__javascript_tool",
"mcp__claude-in-chrome__shortcuts_list",
"mcp__claude-in-chrome__shortcuts_execute",
"mcp__claude-in-chrome__resize_window",
"mcp__claude-in-chrome__switch_browser",
"mcp__claude-in-chrome__upload_image",
"mcp__claude-in-chrome__gif_creator",
"mcp__claude-in-chrome__update_plan",
"mcp__claude-in-chrome__computer"
]
Deny list — irreversible / catastrophic operations that always require confirmation. Note: rm and kill are allowed broadly above; only the genuinely-cannot-be-undone variants live here:
[
"Bash(rm -rf /:*)",
"Bash(rm -rf ~:*)",
"Bash(rm -rf $HOME:*)",
"Bash(git push --force:*)", "Bash(git push -f:*)",
"Bash(git reset --hard:*)", "Bash(git clean -f:*)",
"Bash(chmod 777:*)",
"Bash(shutdown:*)", "Bash(reboot:*)", "Bash(halt:*)", "Bash(poweroff:*)",
"Bash(mkfs:*)", "Bash(dd:*)", "Bash(diskutil erase:*)", "Bash(launchctl:*)"
]
Read ~/.claude/settings.json, merge the allow and deny lists (deduplicating), and write back. Preserve all existing settings — only add to permissions.allow and permissions.deny.
Tell the user what was added and that they need to restart Claude Code for the permissions to take effect.