Security audit

Siluzan CSO

Security checks for vulnerabilities and agentic risk

Overview

The skill’s CSO content and publishing workflow is mostly coherent, but its installer makes broad persistent changes and runs unverified remote code that users should review before installing.

Install only if you trust the Siluzan publisher and need this CSO workflow. Prefer a manual, pinned install over the one-click remote shell or PowerShell commands; review scripts before running them, avoid sudo-piped setup scripts, keep your npm registry unchanged unless you explicitly choose the mirror, and register the skill only into the assistant you intend to use. Before any OAuth, upload, publish, delete, or account-group change, verify the target account and require the dry-run or confirmation steps described in the docs.

A.I.G

Vulnerability Patterns

Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Insecure DependenciesIntroduces malicious components through unsafe dependency sources
Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally

Findings (4)

T03 · Remote Payload Retrieval and Execution

Error

Location: SKILL.md:50
Finding: Mutable Remote Scripts Are Executed Without Integrity Verification
Content: View full analysis
Remediation: View remediation

T03 · Remote Payload Retrieval and Execution

Error

Location: scripts/install.ps1:107
Finding: Unverified Executable Is Downloaded and Silently Launched
Content: View full analysis
Remediation: View remediation

T02 · Agent Memory Poisoning

Error

Location: scripts/install.sh:151
Finding: Installer Forcefully Registers the Skill Across All AI Assistant Profiles
Content: View full analysis
Remediation: View remediation

T08 · Insecure Dependencies

Warning

Location: scripts/install.sh:133
Finding: Installer Persistently Reconfigures the User's npm Registry
Content: View full analysis
/dev/null || echo "") if [ "$current_registry" != "$NPM_MIRROR" ] && [ "$current_registry" != "${NPM_MIRROR}/" ]; then info "Switching npm registry to China mirror for faster downloads..." npm config set registry "$NPM_MIRROR" info "npm registry set to $NPM_MIRROR" else info "npm registry already set to China mirror" fi ``` ```powershell $currentRegistry = '' try { $currentRegistry = (npm config get registry 2>$null).Trim() } catch {} if ($currentRegistry -ne $NPM_MIRROR -and $currentRegistry -ne "$NPM_MIRROR/") { Write-Info 'Switching npm registry to China mirror for faster downloads...' npm config set registry $NPM_MIRROR Write-Info "npm registry set to $NPM_MIRROR" } else { Write-Info 'npm registry already set to China mirror' } ``` ### Technical Analysis The installer uses `npm config set registry`, which persistently changes npm configuration instead of scoping the alternate registry to the installation of `siluzan-cso-cli`. As a result, unrelated npm commands executed after the Skill installation may resolve packages through `https://registry.npmmirror.com`. This expands the supply-chain trust boundary for all future npm activity under the affected configuration. The setting is not restored after installation, and the user is not asked to approve the persistent change. The external `siluzan-cso-cli` package source is not included in the audited project, so its implementation could not be independently reviewed as part of this artifact. ### Attack Path 1. A user runs the installer. 2. The installer detects that npm is configured to use a different registry. 3. It persistently changes the registry to the configured mirro ...[truncated 859 chars]
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration

Findings (164)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 98% confidence
Finding: The skill is presented as a CSO workflow for content, persona, publishing, and RAG tasks, but it also instructs the agent to install software, bootstrap Node.js, register skills into assistant directories, and initiate authentication. This expands from business-task orchestration into host modification and credential setup, increasing attack surface and violating least surprise for a workflow skill.
Content: No source excerpt is available for this finding.

Ae1

High

Category: analysis-evasion
Confidence: 100% confidence
Finding: Referenced artifact was not completely inspected
Content: Scanner excerpt · SKILL.md (reported line 5)May include surrounding context.
md
(1) **文案生成与改稿**：选题、爆款拆解、新写成稿（公众号、小红书、**视频口播/字幕/配音/分镜脚本**、博客、改稿润色、评论区回复等）须走 `three-lib-content-workflow/content-writer.workflow.md`；**热点/资讯生成选题**见 `topic-selecti

Ae1

High

Category: analysis-evasion
Confidence: 100% confidence
Finding: Referenced artifact was not completely inspected
Content: Scanner excerpt · SKILL.md (reported line 89)May include surrounding context.
md
(1) **文案生成与改稿**：选题、爆款拆解、新写成稿（公众号、小红书、**视频口播/字幕/配音/分镜脚本**、博客、改稿润色、评论区回复等）须走 `three-lib-content-workflow/content-writer.workflow.md`；**热点/资讯生成选题**见 `topic-selecti

Ae1

High

Category: analysis-evasion
Confidence: 100% confidence
Finding: Referenced artifact was not completely inspected
Content: Scanner excerpt · SKILL.md (reported line 93)May include surrounding context.
md
(1) **文案生成与改稿**：选题、爆款拆解、新写成稿（公众号、小红书、**视频口播/字幕/配音/分镜脚本**、博客、改稿润色、评论区回复等）须走 `three-lib-content-workflow/content-writer.workflow.md`；**热点/资讯生成选题**见 `topic-selecti

Ae1

High

Category: analysis-evasion
Confidence: 100% confidence
Finding: Referenced artifact was not completely inspected
Content: Scanner excerpt · SKILL.md (reported line 137)May include surrounding context.
md
(1) **文案生成与改稿**：选题、爆款拆解、新写成稿（公众号、小红书、**视频口播/字幕/配音/分镜脚本**、博客、改稿润色、评论区回复等）须走 `three-lib-content-workflow/content-writer.workflow.md`；**热点/资讯生成选题**见 `topic-selecti

Tool Parameter Abuse

High

Category: Tool Misuse
Confidence: 85% confidence
Finding: The documentation explicitly exposes a --no-check-leak option that disables internal-content leakage detection, even though the tool is intended as a pre-delivery safety gate. In this skill context, the validator is supposed to catch disclosure of internal codes, SOPs, and knowledge assets; documenting an easy bypass materially weakens that control and can lead to accidental or intentional release of sensitive internal workflow content.
Content: Scanner excerpt · references/validate-content.md (reported line 52)May include surrounding context.
md
siluzan-cso workflow validate -f article.md --max 800 --strip-markdown # 关闭内部内容泄漏检查（仅看字数） siluzan-cso workflow validate -f draft.md --max 280 --no-check-leak # 追加自定义禁用词（命中即不通过） siluzan-cso workflow validate -f draft.md --forbidden "竞品名,内部代号,占位符"

Tool Parameter Abuse

High

Category: Tool Misuse
Confidence: 88% confidence
Finding: Listing --no-check-leak in the parameter table normalizes and operationalizes bypass of the leakage check, making misuse easier in automated or routine workflows. Because this validator is specifically designed to prevent disclosure of internal asset names, codes, and SOP markers in user-facing content, a documented bypass undermines a security/privacy control rather than a mere quality check.
Content: Scanner excerpt · references/validate-content.md (reported line 73)May include surrounding context.
md
| `--max <n>` | 字数上限（按 `--count-by` 口径）。口播成稿只约束口播正文，不含字幕重点 / 画面建议等成稿块 | | `--count-by <mode>` | 限制口径：`chars-no-space`（不含空白，**默认**）· `chars`（含空白）· `cjk`（汉字）· `words`（英文单词） | | `--strip-markdown` | 统计前去除 markdown 语法与 frontmatter，默认 `false` | | `--no-check-leak` | 跳过内部骨架资产泄漏检查（默认开启检查） | | `--forbidden <words>` | 额外禁用子串，逗号分隔，命中即校验不通过 | | `--json` | JSON 输出（含全部指标、问题列表、`countScope`、`passed`） |

Lp1

High

Category: MCP Least Privilege
Confidence: 95% confidence
Finding: This installer script has direct shell/install capability and performs package installation, process launching, environment/path changes, and external downloads. If the skill framework did not declare that level of execution permission, users and platforms may be exposed to powerful behavior they did not explicitly authorize, which is a real security boundary issue rather than a mere implementation detail.
Content: No source excerpt is available for this finding.

External Script Fetching

High

Category: Supply Chain
Confidence: 99% confidence
Finding: curl -fsSL https://install-node.vercel.app/lts | bash -s -- --yes executes remote code immediately after download, with no checksum or signature verification. This is a classic unsafe installer pattern that can turn an upstream or network compromise into arbitrary code execution on the user's machine.
Content: Scanner excerpt · scripts/install.sh (reported line 61)May include surrounding context.
sh
brew link --overwrite node@22 2>/dev/null || true else info "Installing Node.js LTS via install-node.vercel.app..." curl -fsSL https://install-node.vercel.app/lts | bash -s -- --yes fi ;; linux)

Chaining Abuse

High

Category: Tool Misuse
Confidence: 98% confidence
Finding: Piping curl output directly into sudo -E bash chains network retrieval to privileged shell execution in one step, eliminating any opportunity for review or integrity checking. This is especially dangerous because compromise of the fetched content immediately becomes root-level code execution.
Content: Scanner excerpt · scripts/install.sh (reported line 67)May include surrounding context.
sh
linux) if command -v apt-get >/dev/null 2>&1; then info "Installing Node.js 22.x via NodeSource (apt)..." curl -fsSL https://deb.nodesource.com/setup_22.x | sudo -E bash - sudo apt-get install -y nodejs elif command -v yum >/dev/null 2>&1; then info "Installing Node.js 22.x via NodeSource (yum)..."

Chaining Abuse

High

Category: Tool Misuse
Confidence: 98% confidence
Finding: This is another direct network-to-root execution chain, making exploitation straightforward if the upstream setup script or delivery path is compromised. The one-liner pattern suppresses user scrutiny and greatly amplifies supply-chain impact.
Content: Scanner excerpt · scripts/install.sh (reported line 71)May include surrounding context.
sh
sudo apt-get install -y nodejs elif command -v yum >/dev/null 2>&1; then info "Installing Node.js 22.x via NodeSource (yum)..." curl -fsSL https://rpm.nodesource.com/setup_22.x | sudo -E bash - sudo yum install -y nodejs else info "Installing Node.js LTS via install-node.vercel.app..."

External Script Fetching

High

Category: Supply Chain
Confidence: 99% confidence
Finding: This line repeats the same unsafe remote-script execution pattern on Linux fallback paths, exposing users to arbitrary code execution from a third-party endpoint. Because it is framed as automatic dependency installation, users may not realize they are trusting and running unaudited code.
Content: Scanner excerpt · scripts/install.sh (reported line 75)May include surrounding context.
sh
sudo yum install -y nodejs else info "Installing Node.js LTS via install-node.vercel.app..." curl -fsSL https://install-node.vercel.app/lts | bash -s -- --yes fi ;; gitbash)

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The skill tells the agent to execute remote installer scripts via curl/bash and PowerShell iex, which are fetched at runtime and executed without integrity verification. This is dangerous because any compromise of the distribution endpoint, dependency chain, or latest tag could lead to arbitrary code execution on the host.
Content: No source excerpt is available for this finding.

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 97)May include surrounding context.
md
## 命令索引 | 命令 | 作用 | 详细文档 | | ---------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----------------------------- | | `siluzan-cso login` / `siluzan-cso send-login-code` | 登录 / 配置凭据；手机号 + 验证码两段式登录 | `references/setup.md` | | `siluzan-cso config show/set/clear` | 查看 / 修改 / 清空本地配置 | `references/setup.md` |

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 99)May include surrounding context.
md
| 命令 | 作用 | 详细文档 | | ---------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----------------------------- | | `siluzan-cso login` / `siluzan-cso send-login-code` | 登录 / 配置凭据；手机号 + 验证码两段式登录 | `references/setup.md` | | `siluzan-cso config show/set/clear` | 查看 / 修改 / 清空本地配置 | `references/setup.md` | | `siluzan-cso init` | Skill 文件初始化（写入 AI 助手目录） | `references/setup.md` | | `siluzan-cso update` | 更新 CLI 版本并刷新 Skill 文件 | `references/setup.md` |

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 100)May include surrounding context.
md
| 命令 | 作用 | 详细文档 | | ---------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----------------------------- | | `siluzan-cso login` / `siluzan-cso send-login-code` | 登录 / 配置凭据；手机号 + 验证码两段式登录 | `references/setup.md` | | `siluzan-cso config show/set/clear` | 查看 / 修改 / 清空本地配置 | `references/setup.md` | | `siluzan-cso init` | Skill 文件初始化（写入 AI 助手目录） | `references/setup.md` | | `siluzan-cso update` | 更新 CLI 版本并刷新 Skill 文件 | `references/setup.md` | | `siluzan-cso authorize --media-type <平台>` | 发起媒体账号 OAuth 授权 ...[truncated 25 chars]

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 101)May include surrounding context.
md
| ---------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----------------------------- | | `siluzan-cso login` / `siluzan-cso send-login-code` | 登录 / 配置凭据；手机号 + 验证码两段式登录 | `references/setup.md` | | `siluzan-cso config show/set/clear` | 查看 / 修改 / 清空本地配置 | `references/setup.md` | | `siluzan-cso init` | Skill 文件初始化（写入 AI 助手目录） | `references/setup.md` | | `siluzan-cso update` | 更新 CLI 版本并刷新 Skill 文件 | `references/setup.md` | | `siluzan-cso authorize --media-type <平台>` | 发起媒体账号 OAuth 授权 | `references/authorize.md` | | `siluzan-cso list-accounts` | 列出媒体账号，获取账号 ID / 数据总览 ...[truncated 25 chars]

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 104)May include surrounding context.
md
| ---------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----------------------------- | | `siluzan-cso login` / `siluzan-cso send-login-code` | 登录 / 配置凭据；手机号 + 验证码两段式登录 | `references/setup.md` | | `siluzan-cso config show/set/clear` | 查看 / 修改 / 清空本地配置 | `references/setup.md` | | `siluzan-cso init` | Skill 文件初始化（写入 AI 助手目录） | `references/setup.md` | | `siluzan-cso update` | 更新 CLI 版本并刷新 Skill 文件 | `references/setup.md` | | `siluzan-cso authorize --media-type <平台>` | 发起媒体账号 OAuth 授权 | `references/authorize.md` | | `siluzan-cso list-accounts` | 列出媒体账号，获取账号 ID / 数据总览 ...[truncated 25 chars]

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 103)May include surrounding context.
md
| `siluzan-cso config show/set/clear` | 查看 / 修改 / 清空本地配置 | `references/setup.md` | | `siluzan-cso init` | Skill 文件初始化（写入 AI 助手目录） | `references/setup.md` | | `siluzan-cso update` | 更新 CLI 版本并刷新 Skill 文件 | `references/setup.md` | | `siluzan-cso authorize --media-type <平台>` | 发起媒体账号 OAuth 授权 | `references/authorize.md` | | `siluzan-cso list-accounts` | 列出媒体账号，获取账号 ID / 数据总览 | `references/list-accounts.md` | | `siluzan-cso persona list` | 拉取 CSO 人设列表。**请先阅读详细文档，规范操作，避免误用。** | `references/persona.md` | | `siluzan-cso rag list` | 列出知识库文件夹；`--rag-only` 仅已建索引；`--folder-id` 查指定文件夹下的子库 ...[truncated 25 chars]

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 109)May include surrounding context.
md
| `siluzan-cso config show/set/clear` | 查看 / 修改 / 清空本地配置 | `references/setup.md` | | `siluzan-cso init` | Skill 文件初始化（写入 AI 助手目录） | `references/setup.md` | | `siluzan-cso update` | 更新 CLI 版本并刷新 Skill 文件 | `references/setup.md` | | `siluzan-cso authorize --media-type <平台>` | 发起媒体账号 OAuth 授权 | `references/authorize.md` | | `siluzan-cso list-accounts` | 列出媒体账号，获取账号 ID / 数据总览 | `references/list-accounts.md` | | `siluzan-cso persona list` | 拉取 CSO 人设列表。**请先阅读详细文档，规范操作，避免误用。** | `references/persona.md` | | `siluzan-cso rag list` | 列出知识库文件夹；`--rag-only` 仅已建索引；`--folder-id` 查指定文件夹下的子库 ...[truncated 25 chars]

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 105)May include surrounding context.
md
| `siluzan-cso update` | 更新 CLI 版本并刷新 Skill 文件 | `references/setup.md` | | `siluzan-cso authorize --media-type <平台>` | 发起媒体账号 OAuth 授权 | `references/authorize.md` | | `siluzan-cso list-accounts` | 列出媒体账号，获取账号 ID / 数据总览 | `references/list-accounts.md` | | `siluzan-cso persona list` | 拉取 CSO 人设列表。**请先阅读详细文档，规范操作，避免误用。** | `references/persona.md` | | `siluzan-cso rag list` | 列出知识库文件夹；`--rag-only` 仅已建索引；`--folder-id` 查指定文件夹下的子库 | `references/rag.md` | | `siluzan-cso rag query` | 知识库向量检索；**`--partition wiki` 或 `default`**（默认 `default`；写稿与须贴库作答时优先 **wiki**，不足再 **default**）；`-q` 含空白时多词分检合并；`--folder-id` / `--tags` 见 `references/rag.md` | `references/rag.md` | | `siluzan-cso account-group list/create/add-accounts/remove-accounts/update/delete` | 账号分组管理 ...[truncated 25 chars]

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 106)May include surrounding context.
md
| `siluzan-cso authorize --media-type <平台>` | 发起媒体账号 OAuth 授权 | `references/authorize.md` | | `siluzan-cso list-accounts` | 列出媒体账号，获取账号 ID / 数据总览 | `references/list-accounts.md` | | `siluzan-cso persona list` | 拉取 CSO 人设列表。**请先阅读详细文档，规范操作，避免误用。** | `references/persona.md` | | `siluzan-cso rag list` | 列出知识库文件夹；`--rag-only` 仅已建索引；`--folder-id` 查指定文件夹下的子库 | `references/rag.md` | | `siluzan-cso rag query` | 知识库向量检索；**`--partition wiki` 或 `default`**（默认 `default`；写稿与须贴库作答时优先 **wiki**，不足再 **default**）；`-q` 含空白时多词分检合并；`--folder-id` / `--tags` 见 `references/rag.md` | `references/rag.md` | | `siluzan-cso account-group list/create/add-accounts/remove-accounts/update/delete` | 账号分组管理 | `references/account-group.md` | | `siluzan-cso upload -f <file>` | 上传视频 / 图片到素材库 ...[truncated 25 chars]

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 108)May include surrounding context.
md
| `siluzan-cso persona list` | 拉取 CSO 人设列表。**请先阅读详细文档，规范操作，避免误用。** | `references/persona.md` | | `siluzan-cso rag list` | 列出知识库文件夹；`--rag-only` 仅已建索引；`--folder-id` 查指定文件夹下的子库 | `references/rag.md` | | `siluzan-cso rag query` | 知识库向量检索；**`--partition wiki` 或 `default`**（默认 `default`；写稿与须贴库作答时优先 **wiki**，不足再 **default**）；`-q` 含空白时多词分检合并；`--folder-id` / `--tags` 见 `references/rag.md` | `references/rag.md` | | `siluzan-cso account-group list/create/add-accounts/remove-accounts/update/delete` | 账号分组管理 | `references/account-group.md` | | `siluzan-cso upload -f <file>` | 上传视频 / 图片到素材库 | `references/upload.md` | | `siluzan-cso extract-cover -f <video> -p <平台>` | 从视频截取封面帧 | `references/extract-cover.md` | | `siluzan-cso publish -c config.json` | 提交多平台发布任务 ...[truncated 25 chars]

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 110)May include surrounding context.
md
| `siluzan-cso rag query` | 知识库向量检索；**`--partition wiki` 或 `default`**（默认 `default`；写稿与须贴库作答时优先 **wiki**，不足再 **default**）；`-q` 含空白时多词分检合并；`--folder-id` / `--tags` 见 `references/rag.md` | `references/rag.md` | | `siluzan-cso account-group list/create/add-accounts/remove-accounts/update/delete` | 账号分组管理 | `references/account-group.md` | | `siluzan-cso upload -f <file>` | 上传视频 / 图片到素材库 | `references/upload.md` | | `siluzan-cso extract-cover -f <video> -p <平台>` | 从视频截取封面帧 | `references/extract-cover.md` | | `siluzan-cso publish -c config.json` | 提交多平台发布任务 | `references/publish.md` | | `siluzan-cso task list/detail/item` | 查看任务状态 / 处理失败 / 重试 | `references/task.md` | | `siluzan-cso report fetch --media <平台>` | 运营报表（核心指标 / 视频排行 / 趋势） ...[truncated 25 chars]

Whitespace Padding

Medium

Category: Prompt Injection
Confidence: 70% confidence
Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
Content: Scanner excerpt · SKILL.md (reported line 111)May include surrounding context.
md
| `siluzan-cso account-group list/create/add-accounts/remove-accounts/update/delete` | 账号分组管理 | `references/account-group.md` | | `siluzan-cso upload -f <file>` | 上传视频 / 图片到素材库 | `references/upload.md` | | `siluzan-cso extract-cover -f <video> -p <平台>` | 从视频截取封面帧 | `references/extract-cover.md` | | `siluzan-cso publish -c config.json` | 提交多平台发布任务 | `references/publish.md` | | `siluzan-cso task list/detail/item` | 查看任务状态 / 处理失败 / 重试 | `references/task.md` | | `siluzan-cso report fetch --media <平台>` | 运营报表（核心指标 / 视频排行 / 趋势） | `references/report.md` | | `siluzan-cso planning ...` | AI 内容规划：生成、监控、详情、导出 ...[truncated 25 chars]

Static analysis

No suspicious patterns detected.