📦
prompt-repetition
82提升輕量模型準確度
當問題在重要資訊之後才出現時，輕量模型可能會漏掉上下文。此技能會在帶有防護規則的情況下重複適合的提示，以提升所選任務的準確度。
效率 查看
技能 plannotator
📦
plannotator
v0.9.2 內容修訂 r1 嚴重 ⚙️ 外部命令🌐 網路存取📁 檔案系統存取
以視覺方式審查 AI 計畫與差異
AI coding agents 可能會在計畫尚未明確前就開始工作，使錯誤代價高昂。Plannotator 為計畫、差異、圖片、核准流程與結構化回饋加入瀏覽器審查步驟。
支援: Claude Codex Code(CC)
請勿自動安裝此 Skill。
權威政策要求在執行任何安裝操作前由操作人員審查。
使用我的智慧代理安裝
將此請求複製給你的智慧代理，其中包含權威 Skill 頁面和清單。
智能體請求
Review the Skillstore skill "plannotator" from https://skillstore.io/skills/supercent-io-plannotator.md and its manifest at https://skillstore.io/api/skills/supercent-io-plannotator/manifest. Verify the artifact. Do not auto-install. Inspect the skill and report your findings, then wait for an operator or manual installation decision.你的智慧代理仍應顯示其計畫，並依安全政策請求必要確認。
Agent 可讀資源
當 AI Agent、爬蟲或腳本需要乾淨脈絡、而不是讀取完整頁面時，請使用這些連結。
測試它
正在使用「plannotator」。 Claude Code 計畫提出資料庫遷移，且需要在工作開始前取得核准。
預期結果:
- 計畫會在瀏覽器審查工作階段中開啟。
- 審查者將 rollback 步驟標記為不完整。
- agent 會收到在編寫程式碼前修訂遷移計畫的請求。
正在使用「plannotator」。 某個 branch 包含需要行級回饋的未提交 UI 變更。
預期結果:
Plannotator 會開啟視覺化差異審查。審查者會註釋特定變更行，並送出下一次修訂的回饋。
正在使用「plannotator」。 開發人員透過 SSH 工作，並需要瀏覽器存取審查 UI。
預期結果:
遠端設定工作流程會設定固定 port，並說明如何將審查 UI forward 到本機瀏覽器。
安全審計
嚴重v6 • 2026/7/9 開啟版本化報告
The skill provides a legitimate visual review workflow, but it has high security impact because installation and setup scripts execute or encourage remote code execution and persistently modify agent configuration. Many Markdown backtick and read-only status findings are false positives, while the installer, hidden configuration writes, temporary plan files, and shell profile editing remain confirmed risks.
10
已掃描檔案
2,251
分析行數
90
複核項目
0
已忽略的誤報
已確認的安全問題 (4)
嚴重
Pipe to shell pattern
if curl -fsSL https://plannotator.ai/install.sh | bash; then
The installer actively pipes https://plannotator.ai/install.sh into bash, so remote code executes without pinning or verification.
嚴重
Pipe to shell pattern
echo -e "${YELLOW}Try manual install: curl -fsSL https://plannotator.ai/install.sh | bash${NC}"
The script prints a manual fallback that pipes a remote installer into bash. This is dangerous guidance because users may execute unverified remote code.
高
Shell Profile Injection via Remote Configuration
The remote configuration script appends PLANNOTATOR_PORT, PLANNOTATOR_BROWSER, and PLANNOTATOR_SHARE_URL values from prompts into a shell profile without validation or shell escaping. Crafted values with quotes, newlines, or shell syntax could execute when the profile is sourced.
The script builds export lines from raw prompt input and appends them to the selected shell profile. There is no numeric port validation or robust shell quoting before the write.
高
scripts/setup-hook.sh:64-78scripts/setup-gemini-hook.sh:190-238scripts/setup-codex-hook.sh:71-83scripts/setup-opencode-plugin.sh:129-145
Persistent Agent Instruction and Hook Installation
The setup scripts install persistent hooks, prompts, slash commands, and developer instructions that change future AI agent behavior. This is core functionality, but it gives the skill durable influence over Claude Code, Gemini CLI, Codex, and OpenCode sessions.
The cited scripts explicitly define hook JSON, GEMINI.md text, Codex developer instructions, and OpenCode command files. These changes are written to user or project configuration and persist beyond the setup command.
能力複核項目 (90)
這些是真實的本機能力，對此技能而言可能是預期行為，因此需要複核，但不會按已確認的惡意行為計分。
高
Hidden file access
echo "$HOME/.zshrc"
The script selects a hidden shell profile and later appends plannotator environment variables to it. Persistent profile edits can affect future shells.
高
Hidden file access
if [ -f "$HOME/.bash_profile" ]; then
The script selects a hidden shell profile and later appends plannotator environment variables to it. Persistent profile edits can affect future shells.
高
Hidden file access
echo "$HOME/.bash_profile"
The script selects a hidden shell profile and later appends plannotator environment variables to it. Persistent profile edits can affect future shells.
高
Hidden file access
echo "$HOME/.bashrc"
The script selects a hidden shell profile and later appends plannotator environment variables to it. Persistent profile edits can affect future shells.
高
Hidden file access
echo "$HOME/.profile"
The script selects a hidden shell profile and later appends plannotator environment variables to it. Persistent profile edits can affect future shells.
高
Hardcoded URL
if curl -fsSL https://plannotator.ai/install.sh | bash; then
This URL is fetched by curl and immediately executed by bash during installation. The remote script is not pinned or verified.
高
Hardcoded URL
echo -e "${YELLOW}Try manual install: curl -fsSL https://plannotator.ai/install.sh | bash${NC}"
This line presents an install command that downloads and executes remote code. Even as guidance, it encourages an unsafe installation pattern.
高
Hardcoded URL
echo " irm https://plannotator.ai/install.ps1 | iex"
This line presents an install command that downloads and executes remote code. Even as guidance, it encourages an unsafe installation pattern.
高
Hardcoded URL
echo " curl -fsSL https://plannotator.ai/install.cmd -o install.cmd && install.cmd && del install
This line presents an install command that downloads and executes remote code. Even as guidance, it encourages an unsafe installation pattern.
高
Hidden file in home directory
# 1. Adding developer_instructions to ~/.codex/config.toml
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
# 2. Creating a plannotator prompt at ~/.codex/prompts/plannotator.md
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
echo " 1. Adds plannotator instructions to ~/.codex/config.toml developer_instructions"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
echo " 2. Creates ~/.codex/prompts/plannotator.md for agent role"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
# 1. Adding developer_instructions to ~/.codex/config.toml
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
# 2. Creating a plannotator prompt at ~/.codex/prompts/plannotator.md
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
echo " 1. Adds plannotator instructions to ~/.codex/config.toml developer_instructions"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
echo " 2. Creates ~/.codex/prompts/plannotator.md for agent role"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
CODEX_DIR="$HOME/.codex"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
# Adds the ExitPlanMode hook to ~/.gemini/settings.json
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
# and plannotator instructions to ~/.gemini/GEMINI.md
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
echo " 1. Adds ExitPlanMode hook to ~/.gemini/settings.json"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
echo " 2. Adds plannotator instructions to ~/.gemini/GEMINI.md"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
# PART 1: Hook in ~/.gemini/settings.json
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
# Adds the ExitPlanMode hook to ~/.gemini/settings.json
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
# and plannotator instructions to ~/.gemini/GEMINI.md
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
echo " 1. Adds ExitPlanMode hook to ~/.gemini/settings.json"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
echo " 2. Adds plannotator instructions to ~/.gemini/GEMINI.md"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
GEMINI_DIR="$HOME/.gemini"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
# PART 1: Hook in ~/.gemini/settings.json
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
# Adds the ExitPlanMode hook to ~/.claude/settings.json
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
echo "Adds the plannotator hook to ~/.claude/settings.json"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
# Ensure ~/.claude directory exists
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
# Adds the ExitPlanMode hook to ~/.claude/settings.json
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
echo "Adds the plannotator hook to ~/.claude/settings.json"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
SETTINGS_FILE="$HOME/.claude/settings.json"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
SETTINGS_DIR="$HOME/.claude"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
# Ensure ~/.claude directory exists
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
# and copies slash commands to ~/.config/opencode/command/
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
echo " 2. Copies slash commands to ~/.config/opencode/command/"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
# and copies slash commands to ~/.config/opencode/command/
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
echo " 2. Copies slash commands to ~/.config/opencode/command/"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file access
OPENCODE_COMMAND_DIR="${XDG_CONFIG_HOME:-$HOME/.config}/opencode/command"
This setup script targets hidden AI tool configuration under the user home directory. Those persistent changes can make future agent sessions run plannotator hooks or prompts.
高
Hidden file in home directory
# Add hook to ~/.claude/settings.json
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file in home directory
- Merges `ExitPlanMode` hook into `~/.claude/settings.json` safely (backs up first)
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file in home directory
- Claude Code hook in `~/.claude/settings.json` (or plugin detected)
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file in home directory
- Gemini CLI hook in `~/.gemini/settings.json`
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file in home directory
- Codex CLI `~/.codex/config.toml` developer_instructions
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file in home directory
- Merges `ExitPlanMode` hook into `~/.gemini/settings.json` (same format as Claude Code)
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file in home directory
- Appends plannotator usage instructions to `~/.gemini/GEMINI.md`
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file in home directory
- Adds plannotator instruction to `developer_instructions` in `~/.codex/config.toml`
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file in home directory
- Creates `~/.codex/prompts/plannotator.md` (invoke with `/prompts:plannotator`)
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file access
# Add hook to ~/.claude/settings.json
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file access
- Merges `ExitPlanMode` hook into `~/.claude/settings.json` safely (backs up first)
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file access
- Claude Code hook in `~/.claude/settings.json` (or plugin detected)
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file access
- Gemini CLI hook in `~/.gemini/settings.json`
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file access
- Codex CLI `~/.codex/config.toml` developer_instructions
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file access
- Merges `ExitPlanMode` hook into `~/.gemini/settings.json` (same format as Claude Code)
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file access
- Appends plannotator usage instructions to `~/.gemini/GEMINI.md`
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file access
- Adds plannotator instruction to `developer_instructions` in `~/.codex/config.toml`
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file access
- Creates `~/.codex/prompts/plannotator.md` (invoke with `/prompts:plannotator`)
The documentation instructs users to configure hidden AI tool files for hooks, prompts, or developer instructions. That is a real persistent configuration risk.
高
Hidden file in home directory
scripts/setup-hook.sh — Automatically adds ExitPlanMode hook to ~/.claude/settings.json
The compact manifest documents automatic modification of the Claude Code hidden settings file. That is a persistent agent configuration change.
高
Hidden file access
scripts/setup-hook.sh — Automatically adds ExitPlanMode hook to ~/.claude/settings.json
The compact manifest documents automatic modification of the Claude Code hidden settings file. That is a persistent agent configuration change.
中
Unix shell invocation
#!/bin/bash
This bash entry point performs installation or persistent configuration changes when run. The shell capability is relevant to the skill risk profile.
中
Unix shell invocation
#!/bin/bash
This bash entry point performs installation or persistent configuration changes when run. The shell capability is relevant to the skill risk profile.
中
Ruby/shell backtick execution
PLANNOTATOR_INSTRUCTION='Before implementing any plan, write the plan markdown to /tmp/plan.md and r
The line embeds shell commands into persistent Codex developer instructions. Future Codex sessions may run those commands before implementation.
中
Template literal with command substitution
PLANNOTATOR_INSTRUCTION='Before implementing any plan, write the plan markdown to /tmp/plan.md and r
The line embeds shell commands into persistent Codex developer instructions. Future Codex sessions may run those commands before implementation.
中
Unix shell invocation
#!/bin/bash
This bash entry point performs installation or persistent configuration changes when run. The shell capability is relevant to the skill risk profile.
中
Temp directory access
PLANNOTATOR_INSTRUCTION='Before implementing any plan, write the plan markdown to /tmp/plan.md and r
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Temp directory access
cat > /tmp/plan.md << '"'"'PLAN
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Temp directory access
python3 -c "import json; print(json.dumps({\"tool_input\": {\"plan\": open(\"/tmp/plan.md\").read(),
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Temp directory access
echo -e " ${GREEN}cat > /tmp/plan.md << 'PLAN'${NC}"
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Temp directory access
echo -e " ${GREEN}python3 -c \"import json; print(json.dumps({\\\"tool_input\\\": {\\\"plan\\\":
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Python file write/append
with open(path, "w") as f:
The embedded Python writes or appends configuration files for supported AI tools. Persistent config writes can alter future agent behavior and should require explicit approval.
中
Python file write/append
with open(path, "a") as f:
The embedded Python writes or appends configuration files for supported AI tools. Persistent config writes can alter future agent behavior and should require explicit approval.
中
Unix shell invocation
#!/bin/bash
This bash entry point performs installation or persistent configuration changes when run. The shell capability is relevant to the skill risk profile.
中
Temp directory access
cat > /tmp/plan.md << 'PLAN'
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Temp directory access
print(json.dumps({'tool_input': {'plan': open('/tmp/plan.md').read(), 'permission_mode': 'acceptEdit
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Temp directory access
" | plannotator > /tmp/plannotator_feedback.txt 2>&1 &
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Temp directory access
echo -e " ${GREEN}cat > /tmp/plan.md << 'PLAN'${NC}"
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Temp directory access
echo -e " ${GREEN}python3 -c \"import json; print(json.dumps({'tool_input': {'plan': open('/tmp/pla
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Python file write/append
with open(path, "w") as f:
The embedded Python writes or appends configuration files for supported AI tools. Persistent config writes can alter future agent behavior and should require explicit approval.
中
Unix shell invocation
#!/bin/bash
This bash entry point performs installation or persistent configuration changes when run. The shell capability is relevant to the skill risk profile.
中
Python file write/append
with open(path, "w") as f:
The embedded Python writes or appends configuration files for supported AI tools. Persistent config writes can alter future agent behavior and should require explicit approval.
中
Ruby/shell backtick execution
!`plannotator review`
The line creates OpenCode slash command content that invokes plannotator through command syntax. It is intentional, but it registers persistent command execution behavior.
中
Ruby/shell backtick execution
!`plannotator annotate "$ARGUMENTS"`
The line creates OpenCode slash command content that invokes plannotator through command syntax. It is intentional, but it registers persistent command execution behavior.
中
Unix shell invocation
#!/bin/bash
This bash entry point performs installation or persistent configuration changes when run. The shell capability is relevant to the skill risk profile.
中
Python file write/append
with open(path, "w") as f:
The embedded Python writes or appends configuration files for supported AI tools. Persistent config writes can alter future agent behavior and should require explicit approval.
中
Temp directory access
" | plannotator > /tmp/plannotator_feedback.txt 2>&1 &
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Temp directory access
" | plannotator > /tmp/plannotator_feedback.txt 2>&1 &
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
中
Temp directory access
7,Manual Plan Submit,Write to /tmp/plan.md then python3 json.dumps(tool_input) | plannotator (avoid
The workflow uses predictable /tmp files for plan or feedback content. That can expose sensitive implementation details or collide with other local users.
風險因素
⚙️ 外部命令 (50)
scripts/check-status.sh:18 scripts/check-status.sh:19 scripts/check-status.sh:20 scripts/check-status.sh:32 scripts/check-status.sh:34 scripts/check-status.sh:240 scripts/check-status.sh:243 scripts/check-status.sh:1 scripts/configure-remote.sh:86 scripts/configure-remote.sh:88 scripts/configure-remote.sh:99 scripts/configure-remote.sh:1 scripts/install.sh:57 scripts/install.sh:78 scripts/install.sh:97 scripts/install.sh:138 scripts/install.sh:140 scripts/install.sh:115 scripts/install.sh:1 scripts/review.sh:39 scripts/review.sh:40 scripts/review.sh:45 scripts/review.sh:48 scripts/review.sh:50 scripts/review.sh:1 scripts/setup-codex-hook.sh:71 scripts/setup-codex-hook.sh:145-148 scripts/setup-codex-hook.sh:148-154 scripts/setup-codex-hook.sh:154-164 scripts/setup-codex-hook.sh:164-171 scripts/setup-codex-hook.sh:171-180 scripts/setup-codex-hook.sh:180-186 scripts/setup-codex-hook.sh:186 scripts/setup-codex-hook.sh:187-188 scripts/setup-codex-hook.sh:87 scripts/setup-codex-hook.sh:71-145 scripts/setup-codex-hook.sh:1 scripts/setup-gemini-hook.sh:195-205 scripts/setup-gemini-hook.sh:205-215 scripts/setup-gemini-hook.sh:215-218 scripts/setup-gemini-hook.sh:218-220 scripts/setup-gemini-hook.sh:130 scripts/setup-gemini-hook.sh:1 scripts/setup-hook.sh:122 scripts/setup-hook.sh:1 scripts/setup-opencode-plugin.sh:135 scripts/setup-opencode-plugin.sh:143 scripts/setup-opencode-plugin.sh:82 scripts/setup-opencode-plugin.sh:1 SKILL.md:16
🌐 網路存取 (24)
scripts/check-status.sh:218 scripts/install.sh:51 scripts/install.sh:72 scripts/install.sh:105 scripts/install.sh:110 scripts/install.sh:117 scripts/install.sh:120 scripts/install.sh:125 scripts/install.sh:206 scripts/install.sh:211 scripts/review.sh:88 scripts/setup-codex-hook.sh:193 scripts/setup-opencode-plugin.sh:76 scripts/setup-opencode-plugin.sh:177 SKILL.md:16 SKILL.md:72 SKILL.md:73 SKILL.md:373 SKILL.md:392 SKILL.md:572 SKILL.md:573 SKILL.md:574 SKILL.md:575 SKILL.md:576
📁 檔案系統存取 (50)
scripts/check-status.sh:50 scripts/check-status.sh:60 scripts/check-status.sh:64 scripts/check-status.sh:133 scripts/check-status.sh:140 scripts/check-status.sh:144 scripts/check-status.sh:149 scripts/check-status.sh:162 scripts/check-status.sh:164 scripts/check-status.sh:168 scripts/check-status.sh:173 scripts/check-status.sh:175 scripts/check-status.sh:45 scripts/check-status.sh:50 scripts/check-status.sh:60 scripts/check-status.sh:64 scripts/check-status.sh:69 scripts/check-status.sh:130 scripts/check-status.sh:133 scripts/check-status.sh:140 scripts/check-status.sh:144 scripts/check-status.sh:147 scripts/check-status.sh:149 scripts/check-status.sh:159 scripts/check-status.sh:162 scripts/check-status.sh:164 scripts/check-status.sh:168 scripts/check-status.sh:171 scripts/check-status.sh:173 scripts/check-status.sh:175 scripts/check-status.sh:183 scripts/check-status.sh:186 scripts/check-status.sh:31 scripts/check-status.sh:32 scripts/check-status.sh:49 scripts/check-status.sh:54 scripts/check-status.sh:70 scripts/check-status.sh:114 scripts/check-status.sh:116 scripts/check-status.sh:132 scripts/check-status.sh:134 scripts/check-status.sh:148 scripts/check-status.sh:161 scripts/check-status.sh:187 scripts/check-status.sh:215 scripts/check-status.sh:239 scripts/check-status.sh:240 scripts/check-status.sh:243 scripts/configure-remote.sh:130 scripts/configure-remote.sh:87
偵測到的模式
Pipe to shell pattern×2
審計單位: codex 檢視審計紀錄 →
分享與引用此報告
分享版本化評估報告、中性徽章、嵌入卡片與引用。Skillstore 只報告證據，不判斷該 Skill 是否安全。
複製報告連結
https://skillstore.io/skills/supercent-io-plannotator/audits/6?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown 徽章
[](https://skillstore.io/skills/supercent-io-plannotator?utm_source=security_passport_badge)HTML 徽章
<a href="https://skillstore.io/skills/supercent-io-plannotator?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/supercent-io-plannotator/security.svg" alt="Skillstore security assessment" loading="lazy"></a>嵌入卡片
<iframe src="https://skillstore.io/embed/skills/supercent-io-plannotator.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>學術引用 (APA · BibTeX · CFF)
APA 引用
supercent-io. (2026). plannotator security audit report (audit version 6) [Author version 0.9.2]. Skillstore. https://skillstore.io/skills/supercent-io-plannotator/audits/6BibTeX 引用
@techreport{supercent-io-supercent-io-plannotator-2026,
author = {supercent-io},
title = {plannotator security audit report (audit version 6)},
institution = {Skillstore},
year = {2026},
number = {6},
url = {https://skillstore.io/skills/supercent-io-plannotator/audits/6},
note = {Author version 0.9.2}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "plannotator security audit report (audit version 6)"
version: "0.9.2"
type: report
authors:
- name: "supercent-io"
date-released: "2026-07-09"
url: "https://skillstore.io/skills/supercent-io-plannotator/audits/6"
identifiers:
- type: other
value: "skillstore:supercent-io-plannotator:audit:6"
description: "Skillstore immutable audit report identifier"
Skillstore 評分
評分依據 證據置信度: 中 45
架構
100
可維護性
87
內容
70
社群
87
規範符合度
你可以打造什麼
在編寫程式碼前審查 Agent 計畫
在瀏覽器中開啟提議的實作計畫，註釋有風險的步驟，並在開始變更程式碼前送出結構化回饋。
檢查含行級筆記的 Git Diff
針對未提交的工作或選定的 commit range 啟動視覺化差異審查，然後擷取聚焦的回饋以便修訂。
協調多個 Agent Tools 的審查
在同一本機工作流程中，為 Claude Code、Codex、Gemini CLI 和 OpenCode 設定相似的計畫審查流程。
試試這些提示
開始基本計畫審查
Use plannotator to review my implementation plan before coding. Open the plan review flow and wait for my approval or requested changes.
審查目前的 Git 變更
Open a plannotator diff review for my current git changes. Summarize the review feedback and update the code only after I approve the changes.
設定 Agent 整合
Help me configure plannotator for my AI coding tool. Show the files that will change first, then apply the setup only after I confirm.
執行完整審查工作流程
Create an implementation plan, send it to plannotator for review, revise it from my annotations, then implement only after approval. After implementation, open a diff review and address approved feedback.
最佳實務
- 設定前先執行 dry-run 模式，讓隱藏的設定變更可見。
- 實作前先審查計畫，並在要求變更中包含 acceptance criteria。
- 在多個 AI tools 啟用 hooks 前，先驗證已安裝的指令碼和備份。
避免
- 不要在未獨立驗證的情況下，將遠端安裝程式 pipe 到 shell。
- 不要在確認團隊使用哪些 tools 前啟用所有整合。
- 不要在多使用者系統的共享暫存路徑中放置敏感計畫。
常見問題
這個 skill 可以協助我審查什麼？
它可協助在瀏覽器 UI 中審查 AI agent 計畫、git diffs、markdown annotations，以及相關的實作回饋。
它支援哪些 tools？
此 skill 記錄了 Claude、Claude Code、Codex、Gemini CLI 和 OpenCode 的工作流程。
它會變更我的本機設定嗎？
是。設定指令碼可以在隱藏的 tool configuration directories 下寫入 hooks、prompts 和 instructions。
我可以在設定前預覽變更嗎？
多個設定指令碼包含 dry-run 模式。請在寫入 hook 或 prompt 檔案前使用它們。
它需要網路存取嗎？
安裝程式使用遠端 plannotator.ai 安裝指令碼。執行前請審查並驗證這些指令碼。
它可以將計畫儲存到筆記 app 嗎？
在設定本機 CLI 和 UI 後，文件化的工作流程支援可選的 Obsidian 和 Bear Notes 匯出。
開發者詳細資訊
作者
supercent-io授權
MIT
作者版本
v0.9.2
Skillstore 修訂
r1
引用
3e4b6c31a74a3bd1a291c98cf585d720cb9fbc88
維護新鮮度
2026/7/18
使用量
9 次下載 · 169 次瀏覽
supercent-io 的更多技能
查看全部📦
api-design
80設計更好的 REST 與 GraphQL API
API 團隊需要在實作開始之前建立一致的合約。本技能會引導資源設計、請求格式、錯誤、版本控管與文件撰寫。
程式開發 查看
📦
code-refactoring
79安全地重構程式碼
複雜的程式碼會拖慢變更速度並提高回歸風險。此技能會透過測試、萃取模式和驗證步驟，引導保留行為的重構。
程式開發 查看
📦
user-guide-writing
79建立使用者指南與教學
文件團隊需要清楚的產品說明，讓使用者不需支援也能跟著操作。此技能可為終端使用者規劃入門指南、操作文章、教學與 FAQ 的結構。
文件 查看
📦
debugging
79系統化地除錯程式碼問題
當症狀掩蓋真正原因時，Bug 會浪費工程時間。此技能會引導結構化重現、隔離、根本原因分析、修復與迴歸預防。
程式開發 查看
📦
task-planning
79規劃 Agile 任務與使用者故事
團隊經常收到範圍很廣的功能需求，卻缺少明確任務、優先順序或驗收標準。此技能會將功能想法轉換為使用者故事、Sprint 待辦項目、依賴關係，以及完成定義檢查。
效率 查看
相關技能
查看全部📦
gauntlet
90Review Defects and Verify Root-Cause Repairs
由 aakashh242
Reviews can miss failure conditions or accept fixes without reliable evidence. Gauntlet guides scoped adversarial reviews, authorized repairs, and revision-bound verification within a finite budget.
程式開發 查看
📦
nestjs-code-audit
84Audit NestJS Code with Evidence-Backed Findings
由 amirtaherkhani
NestJS repositories can hide boundary flaws and runtime risks that lint results alone cannot reveal. This skill combines local evidence collection and semantic review into one prioritized report.
程式開發 查看
📦
code-reviewer
83使用結構化品質檢查審查程式碼
由 alirezarezvani
程式碼審查可能會遺漏品質、安全性與可維護性問題。此技能會透過腳本、檢查清單與報告產生流程，引導進行結構化審查。
程式開發 查看
📦
context-degradation
81診斷 AI 代理中的上下文退化
由 ChakshuGautam
長對話和大型提示可能會讓 AI 代理遺失重要資訊，或遵循互相衝突的上下文。此技能可協助 Claude、Codex 和 Claude Code 識別退化模式並選擇緩解策略。
程式開發 查看
📦
coding-agent-pm
81管理 Coding Agent 工作流程
由 kiwi-phantomworks
當任務、範圍與檢查不夠明確時，coding agent 專案可能會偏離方向。此 skill 提供 PM 範本與審查規則，讓 Claude、Codex 和 Claude Code 的工作保持有界且可驗證。
程式開發 查看
📦
code-review
80審查程式碼中的錯誤與安全性問題
由 21pounder
程式碼審查可能會漏掉錯誤、安全漏洞和可維護性問題。此技能會引導 Claude、Codex 和 Claude Code 執行結構化的審查步驟。
程式開發 查看