📦
prompt-repetition
82경량 모델 정확도 향상
경량 모델은 중요한 정보 뒤에 질문이 나올 때 맥락을 놓칠 수 있습니다. 이 스킬은 선택된 작업의 정확도를 높이기 위해 가드레일과 함께 적절한 프롬프트를 반복합니다.
생산성 보기
스킬 jeo
📦
jeo
v1.3.1 콘텐츠 리비전 r1 치명적 ⚙️ 외부 명령어🌐 네트워크 접근📁 파일 시스템 접근🔑 환경 변수
JEO로 AI 에이전트 워크플로 조율하기
복잡한 AI 코딩 세션에는 일관된 계획, 검증, 인수인계 제어가 필요합니다. JEO는 계획 검토, 실행, UI 피드백, 정리를 위한 도구 간 워크플로를 제공합니다.
지원: Claude Codex Code(CC)
이 Skill을 자동으로 설치하지 마세요.
공식 정책에 따라 설치 작업 전에 운영자 검토가 필요합니다.
내 에이전트로 설치
이 요청을 에이전트에 복사하세요. 공식 Skill 페이지와 매니페스트가 포함되어 있습니다.
에이전트 요청
Review the Skillstore skill "jeo" from https://skillstore.io/skills/supercent-io-jeo.md and its manifest at https://skillstore.io/api/skills/supercent-io-jeo/manifest. Verify the artifact. Do not auto-install. Inspect the skill and report your findings, then wait for an operator or manual installation decision.에이전트는 계속해서 계획을 표시하고 보안 정책에서 요구하는 확인을 요청해야 합니다.
에이전트가 읽기 쉬운 리소스
AI 에이전트, 크롤러, 스크립트가 전체 페이지 대신 정리된 컨텍스트만 필요할 때 이 링크를 사용하세요.
테스트해 보기
"jeo" 사용 중입니다. JEO로 로그인 양식 업데이트를 계획하세요.
예상 결과:
- 워크플로가 PLAN 단계로 진입하고 작업 상태를 기록합니다.
- 구현이 시작되기 전에 계획 검토 게이트가 준비됩니다.
- 실행은 승인 또는 피드백을 기다린 뒤 진행됩니다.
"jeo" 사용 중입니다. JEO를 사용해 대시보드 변경을 검증하세요.
예상 결과:
- 구현은 브라우저 검증 도구로 확인됩니다.
- 결과에는 관찰된 UI 상태와 남은 문제가 포함됩니다.
- 검증이 완료된 후 정리 지침이 제공됩니다.
"jeo" 사용 중입니다. UI 피드백을 제출한 후 annotate mode를 실행하세요.
예상 결과:
- 제출된 주석이 나열되고 확인됩니다.
- 각 수정 사항이 적용된 뒤 짧은 요약과 함께 해결됨으로 표시됩니다.
- 제출된 주석이 더 이상 남아 있지 않으면 루프가 중지됩니다.
보안 감사
치명적v5 • 2026. 7. 9. 버전별 보고서 열기
The audit confirms serious install and persistence risks, including a remote installer executed through bash, user-level AI tool hooks, and automatic project source modification. Many markdown, localhost, and status-check matches were false positives, but the remaining confirmed behavior is enough to require manual review before publishing. No direct credential exfiltration intent was found. Static review was capped at 400/850 representative findings; omitted static matches are unconfirmed, so automatic publishing stays disabled until manual review.
19
스캔된 파일
5,560
분석된 줄 수
80
검토 항목
0
오탐 무시됨
확인된 보안 우려 사항 (6)
치명적
Pipe to shell pattern
if run "curl -fsSL https://plannotator.ai/install.sh | bash"; then
This line downloads a remote installer and executes it with bash through the installer helper. That pattern gives remote content immediate code execution during installation.
치명적
[HEURISTIC] DANGEROUS COMBINATION: Code execution + Network + Credential access
This combination is common in credential stealers and RATs
The combination is present through installers, subprocess hooks, local HTTP calls, and user configuration access. I did not find credential exfiltration intent, but the combined capability is a real marketplace risk.
치명적
Prompt Injection Attempt Detected
SKILL.md tells agents, "The following are commands, not descriptions. Execute them in order," and setup-codex.sh installs a prompt that says, "You are now operating in JEO mode." This attempts to steer host-agent behavior and should be treated as untrusted control text.
The files contain direct imperative control text that tells the agent to change mode, execute steps, and continue without user confirmation. This matches the prompt-injection patterns specified for this audit.
높음
[HEURISTIC] SUSPICIOUS COMBINATION: Filesystem + Credentials + Network
This combination could indicate credential harvesting and exfiltration
The skill writes user-level AI tool configs and registers network-backed MCP or local HTTP workflows. The intent appears orchestration, but persistent config plus network access is a real risk.
높음
Persistent User-Level Hook Installation
Setup scripts register hooks and MCP servers in user-level Claude, Codex, and Gemini configuration. Those hooks can run JEO scripts on future agent events outside the original installation command.
The setup scripts explicitly write hook commands and MCP server definitions to hidden user configuration files. The behavior is documented and implemented in multiple platform installers.
높음
Automatic Project Source Modification
The agentation helper installs a package and rewrites a discovered React entry point to mount its component. This can change application source without a user-reviewed diff or project-specific allowlist.
The helper opens and writes the entry file, while SKILL.md states the package and component are auto-installed and auto-injected. This is a clear intent-level project modification risk.
기능 검토 항목 (80)
이 스킬에 대해 예상될 수 있는 실제 로컬 기능이므로 검토가 필요하지만, 확인된 악의적 동작으로는 집계되지 않습니다.
치명적
Hardcoded URL
if run "curl -fsSL https://plannotator.ai/install.sh | bash"; then
This URL is used to download and execute a remote plannotator installer. Remote executable install sources require pinning and verification.
높음
Hidden file in home directory
# Configures: omc plugin, plannotator hook, agentation MCP, jeo workflow in ~/.claude/settings.json
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
# ── 2. Configure ~/.claude/settings.json ─────────────────────────────────────
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
info "Configuring ~/.claude/settings.json..."
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
settings_path = os.path.expanduser("~/.claude/settings.json")
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
# Configures: developer_instructions + agentation MCP in ~/.codex/config.toml + /prompts:jeo
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
# ── 2. Configure ~/.codex/config.toml ────────────────────────────────────────
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
info "Configuring ~/.codex/config.toml..."
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
config_path = os.path.expanduser("~/.codex/config.toml")
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
ok "JEO developer_instructions synced in ~/.codex/config.toml"
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
# discovery of plannotator when it is installed in ~/.local/bin.
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
os.path.expanduser("~/.codex/skills/jeo/scripts/plannotator-plan-loop.sh"),
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
os.path.expanduser("~/.agent-skills/jeo/scripts/plannotator-plan-loop.sh"),
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
config_path = os.path.expanduser("~/.codex/config.toml")
This setup script targets user-level hidden configuration paths and registers hooks or MCP settings. Persistent changes under home directories are security-relevant for an installable skill.
높음
Hidden file in home directory
**Config file**: `~/.claude/settings.json`
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
"command": "python3 ~/.agent-skills/jeo/scripts/claude-plan-gate.py",
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
**agentation MCP config** (`~/.claude/settings.json` or `.claude/mcp.json`):
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
"command": "python3 ~/.agent-skills/jeo/scripts/claude-agentation-submit-hook.py",
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
`bash ~/.agent-skills/jeo/scripts/setup-claude.sh` migrates stale `~/.agent-skills/omg/...` hook pat
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
# - developer_instructions: ~/.codex/config.toml
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
# - prompt file: ~/.codex/prompts/jeo.md
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
# - notify hook: ~/.codex/hooks/jeo-notify.py
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
**agentation MCP config** (`~/.codex/config.toml`):
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
**notify hook** (`~/.codex/hooks/jeo-notify.py`):
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
**`~/.codex/config.toml`** config:
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
notify = ["python3", "~/.codex/hooks/jeo-notify.py"]
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
# - AfterAgent backup hook: ~/.gemini/hooks/jeo-plannotator.sh
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
# - Instructions (MANDATORY loop): ~/.gemini/GEMINI.md
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
**AfterAgent backup hook** (`~/.gemini/settings.json`):
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file in home directory
"command": "bash ~/.gemini/hooks/jeo-plannotator.sh",
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Python subprocess.run
proc = subprocess.run(
The Claude hook directly launches the external plannotator binary during plan review. This is intended behavior, but it creates automatic command execution from a persistent hook.
높음
Python subprocess.run
result = subprocess.run(
The generated Codex hook runs bash or plannotator from a notify path. This is an intended orchestration flow, but it is persistent external command execution.
높음
Python subprocess.run
subprocess.run(["plannotator"], input=payload, stdout=f, stderr=f, text=True, env=get_plannotator_en
The generated Codex hook runs bash or plannotator from a notify path. This is an intended orchestration flow, but it is persistent external command execution.
높음
Python subprocess.run
' subprocess.run(["python3", h, raw], check=False)\n'
The dispatcher executes registered hook scripts from a user-level registry. That creates a persistent fan-out execution path for future Codex notifications.
높음
Hidden file access
**Config file**: `~/.claude/settings.json`
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file access
"command": "python3 ~/.agent-skills/jeo/scripts/claude-plan-gate.py",
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file access
**agentation MCP config** (`~/.claude/settings.json` or `.claude/mcp.json`):
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file access
"command": "python3 ~/.agent-skills/jeo/scripts/claude-agentation-submit-hook.py",
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file access
`bash ~/.agent-skills/jeo/scripts/setup-claude.sh` migrates stale `~/.agent-skills/omg/...` hook pat
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file access
# - developer_instructions: ~/.codex/config.toml
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file access
# - prompt file: ~/.codex/prompts/jeo.md
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file access
# - notify hook: ~/.codex/hooks/jeo-notify.py
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Hidden file access
**agentation MCP config** (`~/.codex/config.toml`):
This documentation instructs users to configure persistent hooks in hidden AI tool directories. It is prose, but it describes the same durable execution path implemented by the setup scripts.
높음
Python file write/append
with open(entry_path, 'w', encoding='utf-8') as f:
This script rewrites the discovered React entry point to inject the Agentation component. Automatic source modification in a user project is security-relevant.
높음
Python file write/append
with open(settings_path, "w") as f:
This setup script writes user-level tool configuration, hook files, or prompt files. These writes can persist behavior across future AI tool sessions.
높음
Python file write/append
with open(config_path, "w") as f:
This setup script writes user-level tool configuration, hook files, or prompt files. These writes can persist behavior across future AI tool sessions.
높음
Python file write/append
with open(config_path, "w") as f:
This setup script writes user-level tool configuration, hook files, or prompt files. These writes can persist behavior across future AI tool sessions.
높음
Python file write/append
with open(feedback_file, "w") as f:
This setup script writes user-level tool configuration, hook files, or prompt files. These writes can persist behavior across future AI tool sessions.
높음
Python file write/append
with open(dispatcher_path, "w") as _f:
This setup script writes user-level tool configuration, hook files, or prompt files. These writes can persist behavior across future AI tool sessions.
높음
Python file write/append
with open(registry_path, "w") as _f:
This setup script writes user-level tool configuration, hook files, or prompt files. These writes can persist behavior across future AI tool sessions.
높음
Python file write/append
with open(config_path, "w") as f:
This setup script writes user-level tool configuration, hook files, or prompt files. These writes can persist behavior across future AI tool sessions.
높음
Python file write/append
with open(settings_path, "w") as f:
This setup script writes user-level tool configuration, hook files, or prompt files. These writes can persist behavior across future AI tool sessions.
높음
Python file write/append
with open(settings_path, "w") as f:
This setup script writes user-level tool configuration, hook files, or prompt files. These writes can persist behavior across future AI tool sessions.
높음
Python file write/append
with open(settings_path, "w") as f:
This setup script writes user-level tool configuration, hook files, or prompt files. These writes can persist behavior across future AI tool sessions.
높음
Python file write/append
with open(settings_path, "w") as f:
This setup script writes user-level tool configuration, hook files, or prompt files. These writes can persist behavior across future AI tool sessions.
높음
Python os file operations
os.makedirs(os.path.dirname(settings_path), exist_ok=True)
The os operations create user-level configuration or hook directories for AI tools. This supports persistence and deserves review.
높음
Python os file operations
os.makedirs(feedback_dir, exist_ok=True)
The os operations create user-level configuration or hook directories for AI tools. This supports persistence and deserves review.
높음
Python os file operations
os.makedirs(runtime_home, exist_ok=True)
The os operations create user-level configuration or hook directories for AI tools. This supports persistence and deserves review.
높음
Python os file operations
os.makedirs(os.path.dirname(dispatcher_path), exist_ok=True)
The os operations create user-level configuration or hook directories for AI tools. This supports persistence and deserves review.
중간
Temp directory access
RUNTIME_ROOT="/tmp/jeo-opencode-${SESSION_KEY}"
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
# bash ${JEO_SKILL_DIR}/scripts/plannotator-plan-loop.sh plan.md /tmp/plannotator_feedback.txt
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
bash .agent-skills/jeo/scripts/plannotator-plan-loop.sh plan.md /tmp/plannotator_feedback.txt 3
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
3. Read /tmp/plannotator_feedback.txt
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
bash .agent-skills/jeo/scripts/plannotator-plan-loop.sh plan.md /tmp/plannotator_feedback.txt 3
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
d = json.load(open('/tmp/plannotator_feedback.txt'))
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
" && echo "PLAN_APPROVED — proceed to EXECUTE" || cat /tmp/plannotator_feedback.txt
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
feedback_dir = f"/tmp/jeo-{_session_key}"
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
runtime_home = f"/tmp/jeo-{_session_key}/.plannotator"
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
echo " writes to: /tmp/plannotator_feedback.txt"
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
# bash .agent-skills/jeo/scripts/plannotator-plan-loop.sh plan.md /tmp/plannotator_feedback.txt 3
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
LOCK_FILE="/tmp/jeo-plannotator-direct.lock"
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
bash "$LOOP_SCRIPT" "$PLAN_FILE" /tmp/plannotator_feedback.txt 3
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
PLANNOTATOR_RUNTIME_HOME="/tmp/jeo-$(python3 -c "import hashlib,os; print(f'/tmp/jeo-{hashlib.md5(os
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
" "$PLAN_FILE" | env HOME="$PLANNOTATOR_RUNTIME_HOME" PLANNOTATOR_HOME="$PLANNOTATOR_RUNTIME_HOME" p
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
bash .agent-skills/jeo/scripts/plannotator-plan-loop.sh plan.md /tmp/plannotator_feedback.txt 3
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
3. /tmp/plannotator_feedback.txt 읽기
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp directory access
"bash .agent-skills/jeo/scripts/plannotator-plan-loop.sh plan.md /tmp/plannotator_feedback.txt 3. "
The scripts use predictable temporary paths for plannotator feedback, locks, or runtime homes. This is common tooling behavior but can create collision or tampering risk.
중간
Temp file creation
import tempfile
The hook creates a predictable temp lock file for concurrency control. This is not malicious, but lock files in shared temp space can be tampered with.
중간
Temp file creation
_LOCK_FILE = Path(tempfile.gettempdir()) / "jeo-plannotator-claude-gate.lock"
The hook creates a predictable temp lock file for concurrency control. This is not malicious, but lock files in shared temp space can be tampered with.
낮음
Python environment access
env = os.environ.copy()
The generated hook copies the full process environment before launching plannotator. This is normal subprocess behavior, but it can expose environment variables to child tools.
위험 요인
⚙️ 외부 명령어 (50)
opencode.json.jeo.bak:18 scripts/check-status.sh:176 scripts/check-status.sh:182 scripts/check-status.sh:183 scripts/check-status.sh:184 scripts/check-status.sh:185 scripts/check-status.sh:186 scripts/check-status.sh:187 scripts/check-status.sh:188 scripts/check-status.sh:257 scripts/claude-agentation-submit-hook.py:34 scripts/claude-agentation-submit-hook.py:100 scripts/claude-plan-gate.py:257 scripts/claude-plan-gate.py:50 scripts/claude-stop-continuation.py:34 scripts/ensure-agentation.sh:13 scripts/ensure-agentation.sh:52 scripts/ensure-agentation.sh:151 scripts/ensure-agentation.sh:234 scripts/ensure-agentation.sh:251 scripts/ensure-plannotator.sh:6 scripts/ensure-plannotator.sh:7 scripts/ensure-plannotator.sh:8 scripts/ensure-plannotator.sh:32 scripts/ensure-plannotator.sh:56 scripts/install.sh:8 scripts/install.sh:9 scripts/install.sh:10 scripts/install.sh:90 scripts/install.sh:92 scripts/plannotator-plan-loop.sh:83 scripts/plannotator-plan-loop.sh:121 scripts/plannotator-plan-loop.sh:184 scripts/plannotator-plan-loop.sh:7 scripts/plannotator-plan-loop.sh:39-41 scripts/plannotator-plan-loop.sh:50 scripts/plannotator-plan-loop.sh:58 scripts/plannotator-plan-loop.sh:222 scripts/run-opencode-safe.sh:12 scripts/setup-claude.sh:17 scripts/setup-claude.sh:43 scripts/setup-codex.sh:457 scripts/setup-codex.sh:477 scripts/setup-codex.sh:572 scripts/setup-codex.sh:165 scripts/setup-codex.sh:167 scripts/setup-codex.sh:172 scripts/setup-codex.sh:174 scripts/setup-codex.sh:175 scripts/setup-codex.sh:177
🌐 네트워크 접근 (50)
opencode.json.jeo.bak:2 opencode.json.jeo.bak:18 references/FLOW.md:114 references/FLOW.md:344 references/FLOW.md:347 references/FLOW.md:351 references/FLOW.md:458 scripts/claude-agentation-submit-hook.py:15 scripts/claude-agentation-submit-hook.py:90 scripts/claude-agentation-submit-hook.py:90 scripts/claude-plan-gate.py:214 scripts/ensure-agentation.sh:17 scripts/install.sh:114 scripts/install.sh:134 scripts/install.sh:146 scripts/install.sh:165 scripts/install.sh:169 scripts/install.sh:223 scripts/plannotator-plan-loop.sh:286 scripts/plannotator-plan-loop.sh:291 scripts/plannotator-plan-loop.sh:303 scripts/plannotator-plan-loop.sh:328 scripts/plannotator-plan-loop.sh:338 scripts/setup-claude.sh:29 scripts/setup-claude.sh:145 scripts/setup-claude.sh:164 scripts/setup-claude.sh:226 scripts/setup-codex.sh:283 scripts/setup-codex.sh:500 scripts/setup-codex.sh:197 scripts/setup-codex.sh:237 scripts/setup-codex.sh:498 scripts/setup-gemini.sh:144 scripts/setup-gemini.sh:308 scripts/setup-gemini.sh:319 scripts/setup-opencode.sh:75 SKILL.md:409 SKILL.md:421 SKILL.md:423 SKILL.md:452 SKILL.md:502 SKILL.md:508 SKILL.md:531 SKILL.md:664 SKILL.md:667 SKILL.md:671 SKILL.md:685 SKILL.md:690 SKILL.md:865 SKILL.md:870
📁 파일 시스템 접근 (50)
opencode.json.jeo.bak:18 opencode.json.jeo.bak:18 references/FLOW.md:117 references/FLOW.md:119 references/FLOW.md:210 references/FLOW.md:211 references/FLOW.md:224 references/FLOW.md:225 references/FLOW.md:435 references/FLOW.md:462 references/FLOW.md:197 references/FLOW.md:210 references/FLOW.md:211 references/FLOW.md:224 references/FLOW.md:225 references/FLOW.md:435 references/FLOW.md:462 references/FLOW.md:410 scripts/check-status.sh:75 scripts/check-status.sh:99 scripts/check-status.sh:102 scripts/check-status.sh:140 scripts/check-status.sh:151 scripts/check-status.sh:63 scripts/check-status.sh:64 scripts/check-status.sh:75 scripts/check-status.sh:88 scripts/check-status.sh:93 scripts/check-status.sh:98 scripts/check-status.sh:99 scripts/check-status.sh:102 scripts/check-status.sh:106 scripts/check-status.sh:124 scripts/check-status.sh:129 scripts/check-status.sh:134 scripts/check-status.sh:140 scripts/check-status.sh:144 scripts/check-status.sh:145 scripts/check-status.sh:151 scripts/check-status.sh:153 scripts/check-status.sh:154 scripts/check-status.sh:162 scripts/check-status.sh:177 scripts/check-status.sh:207 scripts/check-status.sh:21 scripts/check-status.sh:30 scripts/check-status.sh:64 scripts/check-status.sh:80 scripts/check-status.sh:88 scripts/check-status.sh:93
🔑 환경 변수 (19)
scripts/check-status.sh:211 scripts/claude-plan-gate.py:43 scripts/ensure-agentation.sh:167 scripts/ensure-agentation.sh:167 scripts/plannotator-plan-loop.sh:94 scripts/plannotator-plan-loop.sh:139 scripts/plannotator-plan-loop.sh:151 scripts/plannotator-plan-loop.sh:152 scripts/plannotator-plan-loop.sh:153 scripts/plannotator-plan-loop.sh:154 scripts/plannotator-plan-loop.sh:197 scripts/setup-claude.sh:60 scripts/setup-codex.sh:349 scripts/setup-gemini.sh:343 scripts/setup-gemini.sh:344 scripts/setup-opencode.sh:62 scripts/setup-opencode.sh:102 scripts/worktree-cleanup.sh:154 SKILL.md:378
감지된 패턴
Pipe to shell pattern[HEURISTIC] DANGEROUS COMBINATION: Code execution + Network + Credential access
감사자: codex 감사 이력 보기 →
이 보고서 공유 및 인용
버전이 지정된 평가 보고서, 중립 배지, 임베드 카드 및 인용을 공유하세요. Skillstore는 근거를 보고하지만 이 Skill의 안전 여부를 판단하지 않습니다.
보고서 링크 복사
https://skillstore.io/skills/supercent-io-jeo/audits/5?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown 배지
[](https://skillstore.io/skills/supercent-io-jeo?utm_source=security_passport_badge)HTML 배지
<a href="https://skillstore.io/skills/supercent-io-jeo?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/supercent-io-jeo/security.svg" alt="Skillstore security assessment" loading="lazy"></a>임베드 카드
<iframe src="https://skillstore.io/embed/skills/supercent-io-jeo.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>학술 인용 (APA · BibTeX · CFF)
APA 인용
supercent-io. (2026). jeo security audit report (audit version 5) [Author version 1.3.1]. Skillstore. https://skillstore.io/skills/supercent-io-jeo/audits/5BibTeX 인용
@techreport{supercent-io-supercent-io-jeo-2026,
author = {supercent-io},
title = {jeo security audit report (audit version 5)},
institution = {Skillstore},
year = {2026},
number = {5},
url = {https://skillstore.io/skills/supercent-io-jeo/audits/5},
note = {Author version 1.3.1}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "jeo security audit report (audit version 5)"
version: "1.3.1"
type: report
authors:
- name: "supercent-io"
date-released: "2026-07-09"
url: "https://skillstore.io/skills/supercent-io-jeo/audits/5"
identifiers:
- type: other
value: "skillstore:supercent-io-jeo:audit:5"
description: "Skillstore immutable audit report identifier"
Skillstore 점수
이 점수의 이유 증거 신뢰도: 중간 64
아키텍처
100
유지보수성
85
콘텐츠
69
커뮤니티
78
사양 준수
만들 수 있는 것
코딩 워크플로 구조화
코딩 작업을 위한 계획 게이트, 실행 단계, 검증 단계, 정리 단계를 만듭니다.
브라우저 UI 변경 검토
피드백을 해결하기 전에 브라우저 검증을 실행하고 제출된 UI 주석을 수집합니다.
에이전트 설정 표준화
여러 AI 코딩 도구 전반에 유사한 오케스트레이션 동작을 설치합니다.
이 프롬프트를 사용해 보세요
작은 작업 계획하기
먼저 JEO를 사용해 이 작업을 계획한 다음, 구현 전에 계획 승인을 기다리세요: <task>.
UI 변경 검증하기
JEO를 사용해 이 UI 변경을 구현하고, 브라우저에서 페이지를 검증한 뒤, 보이는 결과를 요약하세요.
UI 주석 처리하기
UI 주석을 제출한 후 JEO annotate mode를 사용하세요. 각 항목을 확인하고, 수정 사항을 적용한 뒤, 변경된 내용을 보고하세요.
워크플로 설정 감사하기
이 저장소의 JEO 설정을 검토하세요. 어떤 hooks, prompts, 로컬 서비스, 정리 단계가 활성화되어 있는지 식별하세요.
모범 사례
- 먼저 dry-run으로 설정 스크립트를 실행하고 모든 구성 변경을 검토하세요.
- 로컬 hooks가 예상되는 신뢰할 수 있는 저장소에서만 JEO를 사용하세요.
- localhost 서비스를 비공개로 유지하고 워크플로가 끝나면 중지하세요.
피하기
- 신뢰할 수 없는 저장소에서 전체 설치 프로그램을 실행하지 마세요.
- diff를 검토하지 않고 skill이 애플리케이션 진입점을 자동 수정하도록 두지 마세요.
- 민감한 작업에서 생성된 계획 승인을 사람의 검토를 대체하는 것으로 간주하지 마세요.
자주 묻는 질문
JEO는 무엇을 조율하나요?
여러 AI 코딩 도구 전반에서 계획, 실행, 브라우저 검증, UI 주석 검토, 정리를 조율합니다.
어떤 도구를 지원하나요?
보고서에는 Claude, Claude Code, Codex 지원이 나열되어 있으며, 파일에는 Gemini CLI와 OpenCode 설정도 설명되어 있습니다.
외부 도구를 설치하나요?
예. 여러 설정 경로에서 외부 구성 요소를 위해 npm, npx, plugin commands 또는 installer scripts를 사용합니다.
사용자 구성을 변경하나요?
예. 설정 스크립트는 사용자 홈 디렉터리의 Claude, Codex, Gemini, OpenCode 구성 파일을 쓸 수 있습니다.
네트워크 접근이 필요한가요?
일부 흐름은 로컬 HTTP 서비스, 패키지 레지스트리, 플러그인 마켓플레이스 또는 원격 설치 프로그램 URL을 사용합니다.
자동 마켓플레이스 설치에 안전한가요?
아니요. 감사에서는 수동 검토가 필요한 영구 hooks, 원격 설치 프로그램, 자동 소스 수정을 식별합니다.
개발자 세부 정보
작성자
supercent-io라이선스
MIT
작성자 버전
v1.3.1
Skillstore 리비전
r1
참조
3e4b6c31a74a3bd1a291c98cf585d720cb9fbc88
유지보수 최신성
2026. 7. 22.
사용량
7 다운로드 · 130 조회
supercent-io의 다른 스킬
전체 보기📦
api-design
80더 나은 REST 및 GraphQL API 설계
API 팀은 구현을 시작하기 전에 일관된 계약이 필요합니다. 이 스킬은 리소스 설계, 요청 형식, 오류, 버전 관리 및 문서화를 안내합니다.
코딩 및 개발 보기
📦
code-refactoring
79코드 안전하게 리팩터링하기
복잡한 코드는 변경 속도를 늦추고 회귀 위험을 높입니다. 이 스킬은 테스트, 추출 패턴, 검증 단계를 통해 동작을 보존하는 리팩터링을 안내합니다.
코딩 및 개발 보기
📦
user-guide-writing
79사용자 가이드 및 튜토리얼 작성
문서화 팀에는 사용자가 지원 없이 따라할 수 있는 명확한 제품 도움말이 필요합니다. 이 스킬은 최종 사용자를 위한 온보딩 가이드, 방법 안내 문서, 튜토리얼, FAQ를 구조화합니다.
문서화 보기
📦
debugging
79코드 문제를 체계적으로 디버깅하기
증상이 실제 원인을 가리면 버그는 엔지니어링 시간을 낭비하게 만듭니다. 이 스킬은 구조화된 재현, 격리, 근본 원인 분석, 수정, 회귀 방지를 안내합니다.
코딩 및 개발 보기
📦
task-planning
79애자일 작업 및 사용자 스토리 계획
팀은 명확한 작업, 우선순위 또는 인수 기준 없이 포괄적인 기능 요청을 받는 경우가 많습니다. 이 스킬은 기능 아이디어를 사용자 스토리, 스프린트 백로그 항목, 의존성, Definition of Done 점검 항목으로 전환합니다.
생산성 보기
관련 스킬
전체 보기📦
project-creator
82CCGG 프로젝트 워크스페이스 생성
작성자 DaronVee
프로젝트 설정에서는 공유 운영, 인덱스, 로깅 요구사항이 누락되는 경우가 많습니다. 이 스킬은 템플릿과 검증 체크를 통해 Claude, Codex, Claude Code가 구조화된 CCGG 프로젝트를 생성하도록 안내합니다.
생산성 보기
📦
slash-command-encoder
82AI 워크플로를 위한 슬래시 명령 만들기
작성자 DNYoussef
사용자가 긴 자연어 프롬프트에 의존하면 반복적인 AI 워크플로가 느려집니다. 이 스킬은 라우팅, 검증, 체이닝, 도움말 텍스트를 갖춘 간결한 슬래시 명령을 정의하는 데 도움을 줍니다.
생산성 보기
📦
skills-cli
79CLI에서 Agent Skills 관리
작성자 xixu-me
Agent Skills를 찾고 유지 관리하는 방식은 도구와 머신마다 일관되지 않을 수 있습니다. 이 skill은 에이전트에게 검색, 설치, 업데이트, 제거, 백업 및 복원을 위한 명확한 CLI 워크플로를 제공합니다.
생산성 보기
📦
building-agents
85Claude Code 에이전트 구축
작성자 C0ntr0lledCha0s
신뢰할 수 있는 AI 에이전트를 만들려면 명확한 스키마, 안전한 도구, 반복 가능한 리뷰 단계가 필요합니다. 이 스킬은 Claude Code 워크플로를 위한 에이전트 설계, 검증, 마이그레이션, 유지 관리를 안내합니다.
코딩 및 개발 보기
📦
agent-development
85Claude Code 에이전트 구축
작성자 Dami-Randle
구조 없이 신뢰할 수 있는 Claude Code 에이전트를 설계하는 것은 어렵습니다. 이 스킬은 에이전트 프롬프트, 트리거, 도구, 메타데이터를 위한 템플릿, 검증 규칙, 예시를 제공합니다.
코딩 및 개발 보기
📦
building-commands
84Claude Code 슬래시 명령어 빌드
작성자 C0ntr0lledCha0s
팀에는 호출하고 검토하기 쉬운 반복 가능한 AI 워크플로가 필요합니다. 이 스킬은 Claude Code 슬래시 명령어를 생성, 업데이트, 검증, 문서화하는 데 도움이 됩니다.
코딩 및 개발 보기