Back to skill

Security audit

视频解说脚本

Security checks for vulnerabilities and agentic risk

Overview

This skill is a Chinese video narration workflow that clearly discloses its local file writes and optional text-only MiMo review call.

Install only if you are comfortable with a Chinese-language video scripting workflow that can write project files under work_dir. Do not run review.py, or disable the orchestrated review, when narration, transcripts, planning text, or research context should not be sent to MiMo.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (23)

Tainted flow: 'req' from os.environ.get (line 197, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/lib.py (reported line 198)May include surrounding context.

python
for attempt in range(max_retries):
        try:
            req = urllib.request.Request(endpoint, data=data, headers=headers)
            with urllib.request.urlopen(req, timeout=300) as resp:
                return json.loads(resp.read().decode("utf-8"))
        except urllib.error.HTTPError as e:
            body = _sanitize_api_error(e.read().decode("utf-8", errors="replace"))

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

这是明显的不匹配。声明描述的是一个完整的视频策划与解说文案技能,覆盖策划、写稿、校验、宣发文案及可选联网评审等端到端能力;而实际提供的代码片段只是一个局部工具模块,专门处理旁白块的起止时间、可移动锚点、与相邻块/其他块的避碰以及紧邻衔接判定。虽然这类逻辑可能作为“解说校验/排布”的支撑实现存在于更大系统中,但就该代码片段本身而言,其直接行为与声明的主要用途差距很大,无法代表所宣称的核心功能,因此应判定为描述与代码行为不一致。

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding

The skill explicitly describes capabilities to read/write files, access environment variables, and optionally send content to a remote API, but it declares no tool scope or permission boundaries. That makes the effective privilege set implicit and easier to misuse or invoke in unintended contexts, especially where sensitive work_dir contents or API keys may be present.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
95% confidence
Finding

The skill documents an external transmission path in which review.py sends the full narration, evidence excerpts, transcripts, and planning text to a third-party MiMo chat API. Even though media files are excluded and the behavior is disclosed, this is still a real data exfiltration channel that could leak sensitive or proprietary script content, transcripts, or contextual material if operators enable it without strict review.

Content

Scanner excerpt · SKILL.md (reported line 49)May include surrounding context.

md
`review.py` 发给谁:

- MiMo 的 OpenAI 兼容 chat 接口 `<MIMO_API_URL>/chat/completions`,用 `MIMO_API_KEY` 认证,模型为 `MIMO_MODEL`(默认 `mimo-v2.5`)。未设 `MIMO_API_URL` 时,按量计费 key 发往 `https://api.xiaomimimo.com/v1`,`tp-` 开头的 Token Plan key 发往 `MIMO_TOKEN_PLAN_CLUSTER` 选定的 token-plan 集群(默认 `https://token-plan-cn.xiaomimimo.com/v1`)。
- 未设 `MIMO_API_KEY` 时直接退出,不构造也不发送任何请求。

`review.py` 发送什么(只有文字,不发视频、帧图片或音频文件):

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

This markdown file contains user-facing operational instructions exclusively in Chinese, and there is no indication that the user can choose another language or that the skill is intentionally limited to a Chinese-only audience. Under the policy rule for language/locale, forcing a specific language without opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

The instructions force a single language context for the skill's operation and user-facing guidance. Under the stated policy, imposing a specific language without opt-in or justification can be a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The function formats evidence text with the Chinese label "帧实", and later rendering also uses Chinese-only strings for visible output. This imposes a specific language on users without any opt-in, language selection, or documented region-specific justification, which matches the locale policy violation criteria.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The renderer emits Chinese-only user-facing text such as "无 timeline evidence", "画面", "对白", and "无 context-only evidence". Because the file provides no language selection or opt-in mechanism, it enforces a specific locale in natural-language output across all uses.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This code contains natural-language content in Chinese, including section headers and later user-facing runtime messages, with no indication that the skill supports user language preference or offers an opt-in choice. Per the policy, forcing a specific language or locale without user choice is a natural-language policy violation.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/lib.py (reported line 16)May include surrounding context.

python
# ── 配置 ──────────────────────────────────────────────────────────────

DEFAULT_MIMO_API_URL = "https://api.xiaomimimo.com/v1"
DEFAULT_MIMO_TOKEN_PLAN_CLUSTER = "cn"
MIMO_TOKEN_PLAN_API_URLS = {
    "cn": "https://token-plan-cn.xiaomimimo.com/v1",

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The function includes multiple user-visible log and exception messages only in Chinese, such as rate-limit, authentication, and retry notices. Because the file does not offer a language choice or document a justified locale restriction, this forces a specific language on users and violates the stated policy.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This Python file emits user-facing console narration entirely in Chinese, including labels like "段 N", "原声字幕第 N 条", and status lines such as "narration.json 预检失败". That is a natural-language locale constraint embedded in the skill without any opt-in, fallback, or justification that the tool is intentionally region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This code emits natural-language messages in Chinese for at least some user-facing validation output, while other messages in the file are in English. Forcing a specific language without user opt-in or documented locale scoping matches the language/locale policy violation criteria.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The log message "narration lint:通过" is a fixed Chinese-language user-facing string. The file does not indicate that the skill is region-specific or that users can choose the output language, so this appears to violate the language/locale policy.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The RUBRIC instructs the model in Chinese and explicitly requires the summary and findings schema content to be returned in Chinese, which imposes a specific language/locale behavior. In this file there is no visible opt-in, language choice, or documented region-specific justification for forcing Chinese output.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The file contains user-facing strings in Chinese, such as the error/instruction text at L035, and additional CLI/runtime messages elsewhere. There is no indication that the skill offers a language choice or that the Chinese-only behavior is a documented, justified locale restriction, which can violate language/locale policy expectations.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The docstring states that validation 'never rewrites the agent's text, timing, order or metadata' and then says full and cut_output persist only measured overlaps_speech ownership. In practice, the code later writes a modified narration object back to narration.json in full and cut_output modes, which is a metadata rewrite and contradicts the absolute 'never rewrites ... metadata' claim.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This code emits natural-language error text in Chinese for timeline validation failures. The file also uses other Chinese-only user-facing messages, so the skill effectively forces a specific language/locale without user opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The missing-file failure message is written only in Chinese, which is a natural-language locale constraint visible to users. Under the policy, forcing a specific language without opt-in is a violation unless the restriction is explicitly documented and justified.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

This markdown file explicitly tells the skill to write gathered information into a project file, which is a file-modifying operation. While the behavior is described operationally, there is no clear warning to the user that running this step will create or update local data on disk.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
96% confidence
Finding

The function docstring at L09 is written in Chinese, which imposes a specific language in natural-language content without indicating user choice or a documented locale-specific reason. This matches the policy category for language or locale constraints appearing in code comments or string literals.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The SystemExit message at L019 is hard-coded in Chinese ("缺少 ...;...") with no indication that the skill is Chinese-only or that users can choose their language. This creates a natural-language locale policy issue because the file imposes a specific language on users without opt-in or documented justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The success log message is user-visible natural language and is emitted only in Chinese. This imposes a locale choice on users without any explicit opt-in or documented justification in the file.

Content

No source excerpt is available for this finding.