T08 · Insecure Dependencies
Error
- Location
SKILL.md:607- Finding
Unpinned npm Package Execution Exposes the API Credential to Supply-Chain Risk
- Content
View full analysis
- Remediation
View remediation
Security audit
ai-poster-generator
Security checks for vulnerabilities and agentic risk
Overview
This is mostly a hosted image-generation helper, but it needs review because it sends user content and an API key to external services, advertises watermark removal, and recommends running an unpinned npm CLI.
Install only if you are comfortable sending prompts, uploaded images, and a DeepNLP/OneKey access key to the hosted service. Avoid using the watermark-removal capability, do not submit sensitive or regulated content, and avoid the documented `npx onekey` command unless you pin and verify the package version or run it in a restricted environment with a scoped key.
Vulnerability Patterns
- Insecure DependenciesIntroduces malicious components through unsafe dependency sources
- 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 (1)
Vulnerability Patterns
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
- Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
- MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (46)
Context-Inappropriate Capability
High
- Category
- Not specified by scanner
- Confidence
- 99% confidence
- Finding
The documentation explicitly includes a photo-editor mode that can 'remove watermark,' which is unrelated to poster generation and creates a clear abuse path for copyright circumvention and unauthorized content modification. In the context of an innocuous-seeming poster generator, this hidden high-risk capability is especially dangerous because it normalizes misuse under a misleading description.
- Content
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
md This is a demo to generate an AI Poster using just simple text prompt, the AI poster asset size ratio/width/height can be customized. The final generated words on poster are editable, and the generated AI image poster can be modified also. | Item | Value | | --- | ---- | | prompt | Introduction to Difference between MCPs,Skills CLIs in AI Agents | | Task | Task: Generate an AI Poster of topic `Introduction to Difference between MCPs,Skills and Agents CLIs`, Choose Style of Futuristic and scientific styles, enable word editing on Poster of #Headline,#Content,#Footer Set Poster Asset Size: 4:5, 1080x1350 |
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 92% confidence
- Finding
The skill instructs users to submit prompts and uploaded images to external services, but it does not provide a clear privacy notice explaining that user content is transmitted to third-party endpoints. This can lead to unintentional disclosure of sensitive prompts, proprietary images, or personal data.
- Content
External Transmission
Medium
- Category
- Data Exfiltration
- Confidence
- 90% confidence
- Finding
The documented curl command transmits user-supplied data and an access key to an external endpoint at agent.deepnlp.org. External transmission is expected for a hosted image-generation service, but it still creates confidentiality and governance risk if users are not clearly informed or if secrets and sensitive inputs are sent without safeguards.
- Content
bash export DEEPNLP_ONEKEY_ROUTER_ACCESS=your_access_key curl -X POST "https://agent.deepnlp.org/agent_router" \ -H "Content-Type: application/json" \ -H "X-OneKey: $DEEPNLP_ONEKEY_ROUTER_ACCESS" \ -d '{
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
md ### Request Parameters | Field | Required | Description | |---------------|----------|----------------------------------------------------------------------------------------------------------------------------------------| | prompt | Required | Text, User Input Design Goal, e.g. `Genreate a movie character description of Movie Odyssey` | | images | Required | List of Reference Images that user uploads |
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
md | Field | Required | Description | |---------------|----------|----------------------------------------------------------------------------------------------------------------------------------------| | prompt | Required | Text, User Input Design Goal, e.g. `Genreate a movie character description of Movie Odyssey` | | images | Required | List of Reference Images that user uploads | | onekey | Required | DeepNLP Router OneKey Program | | api_id | Required | user `image_generator` to represent image generator | | template_id | Required | e.g. ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor |
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
md |---------------|----------|----------------------------------------------------------------------------------------------------------------------------------------| | prompt | Required | Text, User Input Design Goal, e.g. `Genreate a movie character description of Movie Odyssey` | | images | Required | List of Reference Images that user uploads | | onekey | Required | DeepNLP Router OneKey Program | | api_id | Required | user `image_generator` to represent image generator | | template_id | Required | e.g. ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | Required | int, e.g. 4 cards of instagram/twitter/xiaohongshu |
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
md | prompt | Required | Text, User Input Design Goal, e.g. `Genreate a movie character description of Movie Odyssey` | | images | Required | List of Reference Images that user uploads | | onekey | Required | DeepNLP Router OneKey Program | | api_id | Required | user `image_generator` to represent image generator | | template_id | Required | e.g. ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | Required | int, e.g. 4 cards of instagram/twitter/xiaohongshu | | asset_size | Required | List of Asset Size Available, default to templates, such as List of width,height,ratio, {"width": 1080, "height": 1080, "ratio": "1:1"} |
Description-Behavior Mismatch
Medium
- Category
- Not specified by scanner
- Confidence
- 96% confidence
- Finding
The skill is presented as an AI poster generator, but the documented supported template IDs broaden it into social-media carousels, presentations, app-store assets, icon design, logo design, and photo editing. This scope drift matters because users and downstream systems may grant trust or permissions based on the narrower manifest description while the skill actually enables materially broader content-generation and editing workflows.
- Content
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
md | template_id | Required | e.g. ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | Required | int, e.g. 4 cards of instagram/twitter/xiaohongshu | | asset_size | Required | List of Asset Size Available, default to templates, such as List of width,height,ratio, {"width": 1080, "height": 1080, "ratio": "1:1"} | | mode | Optional | The design complexity, basic,standard,advanced | | styles | Optional | List, List of Styles Prompt, e.g. deep monochromatic tones,techonology styles | | design_config | Optional | Dict, Detailed Layers of Text, Prompt, Images on the Slides |
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
md | template_id | Required | e.g. ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | Required | int, e.g. 4 cards of instagram/twitter/xiaohongshu | | asset_size | Required | List of Asset Size Available, default to templates, such as List of width,height,ratio, {"width": 1080, "height": 1080, "ratio": "1:1"} | | mode | Optional | The design complexity, basic,standard,advanced | | styles | Optional | List, List of Styles Prompt, e.g. deep monochromatic tones,techonology styles | | design_config | Optional | Dict, Detailed Layers of Text, Prompt, Images on the Slides |
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
md #### Supported Template IDs | Template ID | Template Name | Asset Size | | --- |-------------------------------|-------------------------------------------------------------------------------------------------------------| | ai-poster | AI Poster | Various Asset Size, 1:1,4:5,9:16,etc. | | social-media-carousel | Social Media Carousel | Social Media Carousel Multiple Cards for Twitter/Xiaohongshu/Intagram Various Asset Size, 1:1,4:5,9:16,etc. |
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Context Window Stuffing
Medium
- Category
- Memory Poisoning
- Confidence
- 80% confidence
- Finding
Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.
- Content
md ### Design Config Layers AI Brief | Field | Description | |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | template_id | Template ID, supported values: ai-poster,social-media-carousel,presentation,app-store,icon-designer,logo-designer,photo-editor | | card_count | 4 ...[truncated 25 chars]
Static analysis
No suspicious patterns detected.