📦
История аудитов
background-remover - 6 аудиты
Сравнение версий
Изменения возможностей и находок между проверенными версиями, сначала новые.
| Версия | Дата | Результат | Пункты проверки | Изменение к предыдущей |
|---|---|---|---|---|
| v6 Последняя | 7 июл. 2026 г., 07:41 | 1 подтверждено | 27 | Возможности не изменились |
| v5 | 7 июл. 2026 г., 07:41 | 1 подтверждено | 27 | Возможности не изменились |
| v4 | 30 июн. 2026 г., 22:15 | Подтверждённых находок нет | 3 | Возможности не изменились |
| v3 | 30 июн. 2026 г., 22:15 | Подтверждённых находок нет | 3 | Доступ к сетиДоступ к файловой системеПеременные окруженияВнешние команды |
| v2 | 18 мар. 2026 г., 06:55 | Подтверждённых находок нет | 0 | Возможности не изменились |
| v1 | 17 мар. 2026 г., 16:11 | Подтверждённых находок нет | 0 | Базовая |
Версия аудита 6 Последняя
7 июл. 2026 г., 07:41
The skill is a documented client for the verging.ai background removal service and intentionally performs authenticated network uploads of user images. Most markdown backtick findings are false positives, but the curl workflow, external API calls, API key handling, temporary files, and public result URLs are real security and privacy risks.
1
Просканировано файлов
239
Проанализировано строк
32
Пункты проверки
0
Ложные срабатывания проигнорированы
Подтверждённые проблемы безопасности (1)
Высокий
External Image Upload and Public Result Exposure
The skill instructs agents to upload images to verging.ai storage for processing and return processed images through a public URL. Users may expose private photos if they are not warned before upload or if result URLs are shared.
The processing flow explicitly uploads images to external storage and the privacy section states that result images are returned via a public URL.
Пункты проверки возможностей (27)
Это реальные локальные возможности, которые могут ожидаться для этого навыка, поэтому они требуют проверки, но не считаются подтверждённым вредоносным поведением.
Высокий
Generic API/secret keys
- VERGING_API_KEY
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
primaryEnv: VERGING_API_KEY
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
| --api-key | -k | Your API Key | VERGING_API_KEY env |
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
| VERGING_API_KEY | Your API Key |
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
Authorization: ApiKey <your_api_key>
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys · 3 совпадений
curl -H "Authorization: ApiKey $VERGING_API_KEY" \
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys · 2 совпадений
curl -X POST -H "Authorization: ApiKey $VERGING_API_KEY" \
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
- Replace `$VERGING_API_KEY` with your actual API key or set it as an environment variable
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
- API Key can be passed via --api-key parameter or read from environment variable VERGING_API_KEY
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
- Set it via environment variable: `export VERGING_API_KEY="your_key"`
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Средний
Ruby/shell backtick execution
```bash
This block documents curl commands that send an API key and image data to external endpoints. It is documentation rather than embedded code, but the skill instructs the agent to run these external commands.
Средний
Ruby/shell backtick execution
- Every 5 seconds call `/api/v1/background-removal/jobs?job_ids=xxx` to query status
This range includes the documented workflow to poll remote job URLs and use curl to download the result. The static trigger is a markdown backtick, but the surrounding instructions require external command execution.
Средний
Hardcoded URL
https://verging.ai/api/v1/upload-video
The documented upload-video endpoint starts the upload flow for user images. This creates external transfer of user-provided files.
Средний
Hardcoded URL
# "url": "https://...r2.cloudflarestorage.com/...jpg?X-Amz-...",
The example response contains a presigned Cloudflare R2 upload URL. It indicates that image content is transferred to external object storage.
Средний
Hardcoded URL
# "public_url": "https://img.panpan8.com/background-removal/2026-03-11/xxx.jpg"
The example public_url uses an external image delivery domain. This is a real privacy consideration because processed image results are exposed by URL.
Средний
Hardcoded URL
"https://...presigned-url-from-step-1..."
The documented curl PUT sends the local image to a presigned external URL. Even though the URL is a placeholder, the workflow is an external upload.
Средний
Hardcoded URL
https://verging.ai/api/v1/background-removal/create-job
The documented create-job call sends image metadata and job data to verging.ai. This is a core outbound network action.
Средний
Hardcoded URL
- For remote image URLs (https://example.com/image.jpg), use curl to download
The skill instructs the agent to download remote image URLs with curl. User-supplied URL downloads can contact arbitrary hosts and should be constrained.
Средний
Hardcoded URL
User: /background-removal -i https://example.com/photo.jpg
The example demonstrates processing a remote image URL. This confirms the skill may fetch user-provided network resources.
Средний
Temp directory access
- **Temporary files:** Local temporary files are stored in `/tmp/verging-bg-removal/` and cleaned up
The skill documents storing temporary files under /tmp/verging-bg-removal. A fixed temp directory can expose image data or collide across runs if permissions and cleanup are weak.
Низкий
Hardcoded URL
| VERGING_API_URL | API base URL (default: https://verging.ai/api/v1) |
The skill defines https://verging.ai/api/v1 as the default outbound API base URL. This is expected for the service, but it is real network access.
Низкий
Hardcoded URL
https://verging.ai/api/v1/auth/me
The documented curl command calls the verging.ai auth endpoint with the API key. This is real authenticated network access.
Низкий
Hardcoded URL
"https://verging.ai/api/v1/background-removal/jobs?job_ids=123"
The documented job-status request polls an external API endpoint. This is expected behavior, but it is real network access.
Низкий
Hardcoded URL
https://verging.ai/api/v1/background-removal/jobs
The documented jobs endpoint lists remote processing jobs through the external API. This confirms ongoing network dependency.
Факторы риска
⚙️ Внешние команды (19)
🌐 Доступ к сети (14)
📁 Доступ к файловой системе (1)
🔑 Переменные окружения (13)
Аудитор:: codex
7 июл. 2026 г., 07:41
The skill is a documented client for the verging.ai background removal service and intentionally performs authenticated network uploads of user images. Most markdown backtick findings are false positives, but the curl workflow, external API calls, API key handling, temporary files, and public result URLs are real security and privacy risks.
1
Просканировано файлов
239
Проанализировано строк
32
Пункты проверки
0
Ложные срабатывания проигнорированы
Подтверждённые проблемы безопасности (1)
Высокий
External Image Upload and Public Result Exposure
The skill instructs agents to upload images to verging.ai storage for processing and return processed images through a public URL. Users may expose private photos if they are not warned before upload or if result URLs are shared.
The processing flow explicitly uploads images to external storage and the privacy section states that result images are returned via a public URL.
Пункты проверки возможностей (27)
Это реальные локальные возможности, которые могут ожидаться для этого навыка, поэтому они требуют проверки, но не считаются подтверждённым вредоносным поведением.
Высокий
Generic API/secret keys
- VERGING_API_KEY
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
primaryEnv: VERGING_API_KEY
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
| --api-key | -k | Your API Key | VERGING_API_KEY env |
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
| VERGING_API_KEY | Your API Key |
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
Authorization: ApiKey <your_api_key>
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys · 3 совпадений
curl -H "Authorization: ApiKey $VERGING_API_KEY" \
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys · 2 совпадений
curl -X POST -H "Authorization: ApiKey $VERGING_API_KEY" \
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
- Replace `$VERGING_API_KEY` with your actual API key or set it as an environment variable
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
- API Key can be passed via --api-key parameter or read from environment variable VERGING_API_KEY
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Высокий
Generic API/secret keys
- Set it via environment variable: `export VERGING_API_KEY="your_key"`
The skill requires a verging.ai API key and uses it for authenticated API requests. This is intended behavior, but it is real handling of sensitive credentials.
Средний
Ruby/shell backtick execution
```bash
This block documents curl commands that send an API key and image data to external endpoints. It is documentation rather than embedded code, but the skill instructs the agent to run these external commands.
Средний
Ruby/shell backtick execution
- Every 5 seconds call `/api/v1/background-removal/jobs?job_ids=xxx` to query status
This range includes the documented workflow to poll remote job URLs and use curl to download the result. The static trigger is a markdown backtick, but the surrounding instructions require external command execution.
Средний
Hardcoded URL
https://verging.ai/api/v1/upload-video
The documented upload-video endpoint starts the upload flow for user images. This creates external transfer of user-provided files.
Средний
Hardcoded URL
# "url": "https://...r2.cloudflarestorage.com/...jpg?X-Amz-...",
The example response contains a presigned Cloudflare R2 upload URL. It indicates that image content is transferred to external object storage.
Средний
Hardcoded URL
# "public_url": "https://img.panpan8.com/background-removal/2026-03-11/xxx.jpg"
The example public_url uses an external image delivery domain. This is a real privacy consideration because processed image results are exposed by URL.
Средний
Hardcoded URL
"https://...presigned-url-from-step-1..."
The documented curl PUT sends the local image to a presigned external URL. Even though the URL is a placeholder, the workflow is an external upload.
Средний
Hardcoded URL
https://verging.ai/api/v1/background-removal/create-job
The documented create-job call sends image metadata and job data to verging.ai. This is a core outbound network action.
Средний
Hardcoded URL
- For remote image URLs (https://example.com/image.jpg), use curl to download
The skill instructs the agent to download remote image URLs with curl. User-supplied URL downloads can contact arbitrary hosts and should be constrained.
Средний
Hardcoded URL
User: /background-removal -i https://example.com/photo.jpg
The example demonstrates processing a remote image URL. This confirms the skill may fetch user-provided network resources.
Средний
Temp directory access
- **Temporary files:** Local temporary files are stored in `/tmp/verging-bg-removal/` and cleaned up
The skill documents storing temporary files under /tmp/verging-bg-removal. A fixed temp directory can expose image data or collide across runs if permissions and cleanup are weak.
Низкий
Hardcoded URL
| VERGING_API_URL | API base URL (default: https://verging.ai/api/v1) |
The skill defines https://verging.ai/api/v1 as the default outbound API base URL. This is expected for the service, but it is real network access.
Низкий
Hardcoded URL
https://verging.ai/api/v1/auth/me
The documented curl command calls the verging.ai auth endpoint with the API key. This is real authenticated network access.
Низкий
Hardcoded URL
"https://verging.ai/api/v1/background-removal/jobs?job_ids=123"
The documented job-status request polls an external API endpoint. This is expected behavior, but it is real network access.
Низкий
Hardcoded URL
https://verging.ai/api/v1/background-removal/jobs
The documented jobs endpoint lists remote processing jobs through the external API. This confirms ongoing network dependency.
Факторы риска
⚙️ Внешние команды (19)
🌐 Доступ к сети (14)
📁 Доступ к файловой системе (1)
🔑 Переменные окружения (13)
Аудитор:: codex
30 июн. 2026 г., 22:15
Static analysis found many command, network, environment, filesystem, and weak-cryptography patterns. Review confirms legitimate third-party API behavior for image processing, but the skill sends API keys and images to external services and returns public result URLs. No prompt injection, bundled malware, or confirmed malicious intent was found; weak-cryptography alerts are false positives from Markdown text.
1
Просканировано файлов
239
Проанализировано строк
7
Пункты проверки
2
Ложные срабатывания проигнорированы
Пункты проверки возможностей (3)
Это реальные локальные возможности, которые могут ожидаться для этого навыка, поэтому они требуют проверки, но не считаются подтверждённым вредоносным поведением.
Средний
Third-Party Image Upload and Public Result URL
The skill instructs agents to upload user images to verging.ai and R2 storage, then return processed images through a public URL. This is legitimate for the advertised service, but it creates privacy and data exposure risk for sensitive images.
The documented processing flow explicitly uploads images and stores results behind public URLs. The behavior matches the skill purpose, so the risk is privacy exposure rather than confirmed malicious intent.
Средний
API Key Used in Credentialed Network Requests
The skill requires VERGING_API_KEY or an explicit API key and sends it in Authorization headers for curl requests. This is expected for the service, but copied commands or command-line flags can expose secrets through logs, shell history, or process listings.
The API key requirement and Authorization header examples are explicit. The endpoint is the advertised provider, so this is an operational secret-handling concern rather than evidence of exfiltration.
Средний
User-Provided Remote URLs May Be Fetched
The skill tells agents to download remote image URLs with curl before processing. Without validation guidance, this could fetch unintended internal, private, or untrusted resources from the execution environment.
The remote URL workflow is clearly documented, but no implementation is bundled. The risk depends on how the agent validates URLs before running curl.
Статические ложные срабатывания проигнорированы (2)
Эти статические совпадения были отклонены семантической проверкой или совпадали только со схемными токенами, поэтому они показываются для прозрачности, но не влияют на оценку качества.
Низкий
External Command Alerts Are Mostly Documentation Examples
The static Ruby backtick detections map to fenced command examples and Markdown instructions, not bundled executable code. The skill still asks agents to use curl, so command execution should be user-visible and limited to the documented workflow.
The reviewed file is Markdown and contains no separate script file. Some command execution remains intentional because the instructions tell the agent to call curl.
Низкий
Weak Cryptography Alerts Are False Positives
The reported weak-cryptography locations contain descriptive Markdown, tables, or skill metadata. No evidence of MD5, SHA1, DES, RC4, or other cryptographic implementation was found in the reviewed skill file.
The cited lines contain no cryptographic code or algorithm selection. This is a strong false positive from keyword matching in prose.
Факторы риска
🌐 Доступ к сети (8)
📁 Доступ к файловой системе (3)
🔑 Переменные окружения (7)
⚙️ Внешние команды (7)
Обнаруженные паттерны
Credentialed curl Requests to External APIUpload to Presigned Storage and Public Image URLsRemote Image URL Download Through curl
Аудитор:: codex
30 июн. 2026 г., 22:15
Static analysis found many command, network, environment, filesystem, and weak-cryptography patterns. Review confirms legitimate third-party API behavior for image processing, but the skill sends API keys and images to external services and returns public result URLs. No prompt injection, bundled malware, or confirmed malicious intent was found; weak-cryptography alerts are false positives from Markdown text.
1
Просканировано файлов
239
Проанализировано строк
7
Пункты проверки
2
Ложные срабатывания проигнорированы
Пункты проверки возможностей (3)
Это реальные локальные возможности, которые могут ожидаться для этого навыка, поэтому они требуют проверки, но не считаются подтверждённым вредоносным поведением.
Средний
Third-Party Image Upload and Public Result URL
The skill instructs agents to upload user images to verging.ai and R2 storage, then return processed images through a public URL. This is legitimate for the advertised service, but it creates privacy and data exposure risk for sensitive images.
The documented processing flow explicitly uploads images and stores results behind public URLs. The behavior matches the skill purpose, so the risk is privacy exposure rather than confirmed malicious intent.
Средний
API Key Used in Credentialed Network Requests
The skill requires VERGING_API_KEY or an explicit API key and sends it in Authorization headers for curl requests. This is expected for the service, but copied commands or command-line flags can expose secrets through logs, shell history, or process listings.
The API key requirement and Authorization header examples are explicit. The endpoint is the advertised provider, so this is an operational secret-handling concern rather than evidence of exfiltration.
Средний
User-Provided Remote URLs May Be Fetched
The skill tells agents to download remote image URLs with curl before processing. Without validation guidance, this could fetch unintended internal, private, or untrusted resources from the execution environment.
The remote URL workflow is clearly documented, but no implementation is bundled. The risk depends on how the agent validates URLs before running curl.
Статические ложные срабатывания проигнорированы (2)
Эти статические совпадения были отклонены семантической проверкой или совпадали только со схемными токенами, поэтому они показываются для прозрачности, но не влияют на оценку качества.
Низкий
External Command Alerts Are Mostly Documentation Examples
The static Ruby backtick detections map to fenced command examples and Markdown instructions, not bundled executable code. The skill still asks agents to use curl, so command execution should be user-visible and limited to the documented workflow.
The reviewed file is Markdown and contains no separate script file. Some command execution remains intentional because the instructions tell the agent to call curl.
Низкий
Weak Cryptography Alerts Are False Positives
The reported weak-cryptography locations contain descriptive Markdown, tables, or skill metadata. No evidence of MD5, SHA1, DES, RC4, or other cryptographic implementation was found in the reviewed skill file.
The cited lines contain no cryptographic code or algorithm selection. This is a strong false positive from keyword matching in prose.
Факторы риска
🌐 Доступ к сети (8)
📁 Доступ к файловой системе (3)
🔑 Переменные окружения (7)
⚙️ Внешние команды (7)
Обнаруженные паттерны
Credentialed curl Requests to External APIUpload to Presigned Storage and Public Image URLsRemote Image URL Download Through curl
Аудитор:: codex
18 мар. 2026 г., 06:55
All static analysis findings are false positives from pattern matching against documentation examples in SKILL.md. The file contains only Markdown documentation with bash code examples showing API usage. No executable code is present. External commands are documentation examples, network calls reference legitimate API endpoints, and environment variable access is for documented API key configuration. This is a legitimate skill providing CLI interface to verging.ai background removal service.
1
Просканировано файлов
239
Проанализировано строк
0
Пункты проверки
0
Ложные срабатывания проигнорированы
В этом завершенном аудите не зафиксировано подтвержденных проблем безопасности.
Аудитор:: claude
17 мар. 2026 г., 16:11
Static analysis detected 58 potential issues, all of which are false positives from documentation content. The 'external_commands' patterns are markdown code blocks showing curl examples, not actual code execution. Network patterns are legitimate API endpoint URLs in documentation. Environment variable access is standard API key configuration. No executable code or security risks found - this is a skill documentation file providing API usage instructions.
1
Просканировано файлов
239
Проанализировано строк
0
Пункты проверки
0
Ложные срабатывания проигнорированы
В этом завершенном аудите не зафиксировано подтвержденных проблем безопасности.
Аудитор:: claude