Back to skill

Security audit

爱奇艺

Security checks for vulnerabilities and agentic risk

Overview

The iQiyi skill mostly matches its video-search and playback purpose, but it needs Review because optional Authorization tokens can be exposed or sent outside the intended service.

Review before installing if you plan to use authenticated history or favorites. Avoid passing Authorization tokens until the skill redacts request metadata, prevents token output, and restricts authenticated requests to the intended iQiyi HTTPS endpoint. For unauthenticated search/playback use, expect live requests to iQiyi and confirm any qips client launch before execution.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/iqiyi-skill-catalog.mjs:387
Finding

Authorization Header Can Be Forwarded to an Arbitrary Network Destination

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/iqiyi-cli.mjs:135
Finding

Authorization Token Is Retained in Results and Printed by the CLI

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (27)

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 100)May include surrounding context.

md
- 运行时清单来自 [`scripts/iqiyi-skill-catalog.mjs`](scripts/iqiyi-skill-catalog.mjs) 的 `getRecommendStyleCatalog()`。

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 138)May include surrounding context.

md
本 skill 的 qips golden helper 位于 [`scripts/qips-build.mjs`](scripts/qips-build.mjs),对应测试位于 [`scripts/qips-build.test.mjs`](scripts/qips-build.test.mjs)。详细场景路由:

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
85% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · docs/marketplace-distribution.md (reported line 55)May include surrounding context.

每次发布前从真源复制这些目录:

bash
rm -rf dist/agent-skills/iqiyi-skill
mkdir -p dist/agent-skills
cp -R .cursor/skills/iqiyi-skill dist/agent-skills/iqiyi-skill

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
85% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · docs/marketplace-distribution.md (reported line 281)May include surrounding context.

准备包:

bash
rm -rf dist/openclaw/iqiyi-skill
mkdir -p dist/openclaw
cp -R .cursor/skills/iqiyi-skill dist/openclaw/iqiyi-skill

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
85% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · docs/marketplace-distribution.md (reported line 390)May include surrounding context.

准备 npm 包:

bash
rm -rf dist/npm/iqiyi-skill
mkdir -p dist/npm
cp -R .cursor/skills/iqiyi-skill dist/npm/iqiyi-skill

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The file consistently presents the skill description and operating instructions only in Chinese, which can amount to a forced language choice without user opt-in. The policy allows locale constraints only when they are explicitly justified or when users are offered a language option, neither of which is documented here.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill explicitly states that operations and CLI commands perform real network requests by default, but it does not pair that behavior with a clear user-facing notice about what data will be transmitted, to which service, and under what trigger conditions. In an agent setting, this can cause user queries or content identifiers to be sent to an external service without sufficiently explicit consent or transparency, increasing privacy and data-handling risk.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill enables implicit invocation without any stated trigger constraints, exclusions, or user-intent boundaries. That increases the chance the agent will call the skill in loosely related contexts, which can cause unintended external actions such as content search, recommendation, or playback flow initiation without sufficiently explicit user intent.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · docs/marketplace-distribution.md (reported line 56)May include surrounding context.

bash
rm -rf dist/agent-skills/iqiyi-skill
mkdir -p dist/agent-skills
cp -R .cursor/skills/iqiyi-skill dist/agent-skills/iqiyi-skill

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · docs/marketplace-distribution.md (reported line 56)May include surrounding context.

bash
rm -rf dist/agent-skills/iqiyi-skill
mkdir -p dist/agent-skills
cp -R .cursor/skills/iqiyi-skill dist/agent-skills/iqiyi-skill

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
80% confidence
Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · docs/qips-capabilities-feishu-wiki.md (reported line 5)May include surrounding context.

md
本文说明 **`qips://` / `qisu://` 协议**在爱奇艺 PC 客户端(PCA)及内嵌 Web(`@iqiyi/qysdk`)中可实现的**页面跳转、片库与频道上下文、点播/直播/第三方播放、播控与按标题开播**等能力。协议由客户端 `PlayInfo`、`Launcher`(含 `WebLauncher` / `PCALauncher`)解析并执行。

更细的拼写示例、与源码行号对齐的说明见仓库内 **`.cursor/skills/iqiyi-skill/SKILL.md`** 及 **`references/qips/`** 目录下参考文档。

---

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
80% confidence
Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · docs/qips-capabilities-feishu-wiki.md (reported line 210)May include surrounding context.

md
本文说明 **`qips://` / `qisu://` 协议**在爱奇艺 PC 客户端(PCA)及内嵌 Web(`@iqiyi/qysdk`)中可实现的**页面跳转、片库与频道上下文、点播/直播/第三方播放、播控与按标题开播**等能力。协议由客户端 `PlayInfo`、`Launcher`(含 `WebLauncher` / `PCALauncher`)解析并执行。

更细的拼写示例、与源码行号对齐的说明见仓库内 **`.cursor/skills/iqiyi-skill/SKILL.md`** 及 **`references/qips/`** 目录下参考文档。

---

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The documentation explicitly instructs callers to forward an Authorization header to an external service endpoint, but it does not require explicit user consent, warn that credentials will leave the local agent boundary, or constrain what tokens are acceptable to forward. In an agent skill context, this can cause unintended credential disclosure if the agent reuses ambient bearer tokens or blindly propagates secrets to a third-party API based on tool documentation alone.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The documentation explicitly recommends bypassing the SDK and invoking an OS-level deeplink handler directly via Start-Process "qips://...". In a skill whose stated scope is in-skill video navigation/playback, this broadens the trust boundary from app-level routing to host-level process launching, which can enable unsafe external protocol invocation if untrusted qips strings are ever propagated into that guidance or copied into implementations.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

This markdown file defines multiple matchable paths using broad everyday terms, including show, game, live, and hot, without stating activation boundaries or exclusion conditions. Such generic aliases can cause unintended matches if this table is used to drive skill invocation or URL-to-channel routing.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

User-facing natural-language fields such as name and description are consistently Chinese-only, with no indication that users can select another language or locale. This effectively forces a specific language experience and matches the policy category for language or locale constraints without opt-in.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The executeOperation function sends request bodies to a remote IQIYI endpoint and may include an Authorization header, but this execution path has no confirmation prompt, print/log disclosure, or inline warning near the network transmission. Although the catalog metadata mentions authorization pass-through, the actual code path that performs the transmission provides no user-visible notice before sending data off-host.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

This markdown file presents all skill coverage prompts and operational guidance only in Chinese. Under the policy rule for language/locale, forcing a single language without user opt-in or a documented region-specific justification can be a natural-language policy violation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The document explicitly instructs consumers to launch the custom qips:// protocol via system handlers such as Start-Process or open, which can trigger an external desktop application from web or shell contexts. In this skill's context, that behavior is intentional product functionality, but without an explicit warning, consent requirement, or trust boundary guidance, integrators may invoke it unsafely and create unexpected app-launch or deep-link abuse paths.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

该文件以中文撰写并明确面向开发者,但没有说明这是仅适用于中文团队或特定区域的文档,也没有提供可选语言。按照语言/locale 政策,若未提供用户选择或合理限定,强制单一语言属于自然语言层面的合规风险。

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

The file presents all operational guidance and labels exclusively in Chinese, and does not indicate that the skill is region-specific or that users can opt into this language. This may conflict with language/locale policy requirements where user-facing skill materials should provide choice or justified constraints.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

The markdown instructs the skill to return a direct Windows executable download link and includes it in the user-facing fallback template, but it does not warn users that they are being sent to download and install software or advise verifying the source. Because this file is a markdown SOP, user-facing warnings about actions affecting system integrity should be documented here.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The usage string explicitly documents video play as # prints qips, no network, presenting it as a purely local operation. While this file routes video.play through buildLocalQipsResult, that function calls playByTitle6 from another module, so the comment asserts a stronger guarantee than this file itself establishes; the documentation claims absence of network activity rather than merely describing the intended path.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

This code invokes a child process via execFileSync to run a Node script, which is a safety-relevant operation under the code-file warning criteria. In this file there is no confirmation, logging, or comment/docstring explaining that direct execution of another script will occur.

Content

No source excerpt is available for this finding.