История аудитов

SKILL.md:17-26 SKILL.md:137-144 SKILL.md:148-154 SKILL.md:158-165 SKILL.md:169-175 SKILL.md:206-211 SKILL.md:262-267

Ruby/shell backtick execution

> **Install the belt CLI skill:** `npx skills add belt-sh/cli`

The line recommends running npx to install a third-party CLI skill. That can download and execute external package code during setup, so it is a real supply-chain risk.

Средний

Ruby/shell backtick execution · 7 совпадений

```bash

The fenced example runs belt login and belt app run against a remote model service. The arguments are static, but execution requires authentication and sends prompt data to an external CLI.

Низкий

Hardcoded URL

> Requires inference.sh CLI (`belt`). [Install instructions](https://raw.githubusercontent.com/infer

The line links to mutable raw GitHub install instructions for a required external CLI. This is a low-severity supply-chain concern because users may follow unaudited setup content.

Факторы риска

⚙️ Внешние команды (20)

🌐 Доступ к сети (2)

SKILL.md:11-15 SKILL.md:21-25 SKILL.md:171-174 SKILL.md:206-210

Аудитор:: codex

Версия аудита 4

9 июл. 2026 г., 11:38

Most static command hits are Markdown fences, inline CLI names, or documentation examples rather than dynamic shell execution. Confirmed risks are limited to third-party CLI installation guidance, belt commands that call external generation services, and remote processing of prompts or app imagery.

1

Просканировано файлов

271

Проанализировано строк

12

Пункты проверки

0

Ложные срабатывания проигнорированы

Подтверждённые проблемы безопасности (1)

Средний

Third-Party Asset Processing Through belt

The skill directs users to create screenshots and videos through the inference.sh belt CLI. Prompts and referenced app images may be processed by an external service, so private app assets need explicit user approval.

The documentation explicitly depends on inference.sh and shows belt app run commands for remote image, stitching, and video generation. The risk is privacy exposure, not evidence of malicious intent.

Пункты проверки возможностей (9)

Это реальные локальные возможности, которые могут ожидаться для этого навыка, поэтому они требуют проверки, но не считаются подтверждённым вредоносным поведением.

Средний

SKILL.md:17-26 SKILL.md:137-144 SKILL.md:148-154 SKILL.md:158-165 SKILL.md:169-175 SKILL.md:206-211 SKILL.md:262-267

Ruby/shell backtick execution

> **Install the belt CLI skill:** `npx skills add belt-sh/cli`

The line recommends running npx to install a third-party CLI skill. That can download and execute external package code during setup, so it is a real supply-chain risk.

Средний

Ruby/shell backtick execution · 7 совпадений

```bash

The fenced example runs belt login and belt app run against a remote model service. The arguments are static, but execution requires authentication and sends prompt data to an external CLI.

Низкий

Hardcoded URL

> Requires inference.sh CLI (`belt`). [Install instructions](https://raw.githubusercontent.com/infer

The line links to mutable raw GitHub install instructions for a required external CLI. This is a low-severity supply-chain concern because users may follow unaudited setup content.

Факторы риска

⚙️ Внешние команды (20)

🌐 Доступ к сети (2)

Аудитор:: codex

Версия аудита 3

5 июл. 2026 г., 16:13

Several static detections are real low-to-medium risks because the skill asks users to install and run the third-party belt CLI. The remaining detections are Markdown fences, diagrams, or metadata false positives; no prompt injection or covert exfiltration intent was found.

1

Просканировано файлов

271

Проанализировано строк

13

Пункты проверки

0

Ложные срабатывания проигнорированы

Пункты проверки возможностей (11)

Это реальные локальные возможности, которые могут ожидаться для этого навыка, поэтому они требуют проверки, но не считаются подтверждённым вредоносным поведением.

Средний

SKILL.md:17-26 SKILL.md:137-144 SKILL.md:148-154 SKILL.md:158-165 SKILL.md:169-175 SKILL.md:206-211 SKILL.md:262-267

Ruby/shell backtick execution

> **Install the belt CLI skill:** `npx skills add belt-sh/cli`

The inline command asks users to run npx skills add belt-sh/cli, which installs an external skill through a package runner. It is not automatic execution, but it creates a real supply-chain risk if followed.

Средний

Ruby/shell backtick execution · 7 совпадений

```bash

The quick-start bash block instructs belt login and belt app run, which launch a third-party CLI and remote model call. The command family is constrained, but it still executes external tooling.

Средний

SKILL.md:267-269

Ruby/shell backtick execution

```

The section includes a belt app store command after the fenced block. It is a documented CLI action, not automatic execution, but it invokes third-party tooling.

Низкий

SKILL.md:11

Hardcoded URL

Create app store screenshots and preview videos via [inference.sh](https://inference.sh) CLI.

The skill explicitly depends on the inference.sh network service for screenshot and video generation. This is expected, but users should know prompt and asset data may leave their environment.

Низкий

Hardcoded URL

> Requires inference.sh CLI (`belt`). [Install instructions](https://raw.githubusercontent.com/infer

The install instructions use a raw GitHub URL from an external repository. It is a normal documentation link, but it is still a third-party network dependency.

Факторы риска

⚙️ Внешние команды (20)

🌐 Доступ к сети (2)

Аудитор:: codex

Версия аудита 2

5 июл. 2026 г., 16:13

Several static detections are real low-to-medium risks because the skill asks users to install and run the third-party belt CLI. The remaining detections are Markdown fences, diagrams, or metadata false positives; no prompt injection or covert exfiltration intent was found.

1

Просканировано файлов

271

Проанализировано строк

13

Пункты проверки

0

Ложные срабатывания проигнорированы

Пункты проверки возможностей (11)

Это реальные локальные возможности, которые могут ожидаться для этого навыка, поэтому они требуют проверки, но не считаются подтверждённым вредоносным поведением.

Средний

SKILL.md:17-26 SKILL.md:137-144 SKILL.md:148-154 SKILL.md:158-165 SKILL.md:169-175 SKILL.md:206-211 SKILL.md:262-267

Ruby/shell backtick execution

> **Install the belt CLI skill:** `npx skills add belt-sh/cli`

The inline command asks users to run npx skills add belt-sh/cli, which installs an external skill through a package runner. It is not automatic execution, but it creates a real supply-chain risk if followed.

Средний

Ruby/shell backtick execution · 7 совпадений

```bash

The quick-start bash block instructs belt login and belt app run, which launch a third-party CLI and remote model call. The command family is constrained, but it still executes external tooling.

Средний

SKILL.md:267-269

Ruby/shell backtick execution

```

The section includes a belt app store command after the fenced block. It is a documented CLI action, not automatic execution, but it invokes third-party tooling.

Низкий

SKILL.md:11

Hardcoded URL

Create app store screenshots and preview videos via [inference.sh](https://inference.sh) CLI.

The skill explicitly depends on the inference.sh network service for screenshot and video generation. This is expected, but users should know prompt and asset data may leave their environment.

Низкий

Hardcoded URL

> Requires inference.sh CLI (`belt`). [Install instructions](https://raw.githubusercontent.com/infer

The install instructions use a raw GitHub URL from an external repository. It is a normal documentation link, but it is still a third-party network dependency.

Факторы риска

⚙️ Внешние команды (20)

🌐 Доступ к сети (2)

SKILL.md:139-143 SKILL.md:171-174 SKILL.md:208-210

Аудитор:: codex

Версия аудита 1

3 июл. 2026 г., 13:12

The skill is a design guide, but it directs users and agents to install and run external CLI commands. Most static hits are markdown false positives; confirmed risks relate to npx installs, belt remote inference calls, and possible upload of app assets.

1

Просканировано файлов

271

Проанализировано строк

14

Пункты проверки

0

Ложные срабатывания проигнорированы

Подтверждённые проблемы безопасности (1)

Средний

Remote Processing of App Assets

The skill guides users to send prompts and screenshot image filenames through the belt CLI to remote generation or stitching services. Proprietary app screens or unreleased UI assets could be exposed if users run these commands without reviewing data handling.

The examples directly invoke remote model apps and one example references local screenshot filenames. The risk depends on what the user supplies, so confidence is high for privacy exposure but not malicious intent.

Пункты проверки возможностей (11)

Это реальные локальные возможности, которые могут ожидаться для этого навыка, поэтому они требуют проверки, но не считаются подтверждённым вредоносным поведением.

Средний

SKILL.md:17-26 SKILL.md:137-144 SKILL.md:148-154 SKILL.md:158-165 SKILL.md:169-175 SKILL.md:206-211 SKILL.md:262-267

Ruby/shell backtick execution

> **Install the belt CLI skill:** `npx skills add belt-sh/cli`

Line 7 instructs users to run an npx skill installation command. That is a real external command and a supply-chain risk if executed without verification.

Средний

Ruby/shell backtick execution · 7 совпадений

```bash

Lines 17-26 contain bash examples for belt login and belt app run. These commands invoke an external CLI and send prompt data to a remote inference service.

Средний

SKILL.md:267-269

Ruby/shell backtick execution

```

Line 269 tells users to run belt app store. This is a real external CLI command, although it appears to browse available applications rather than process user data.

Низкий

SKILL.md:11

Hardcoded URL

Create app store screenshots and preview videos via [inference.sh](https://inference.sh) CLI.

Line 11 names inference.sh as the service used to create screenshots and videos. This is a legitimate hardcoded external service dependency that should be disclosed.

Низкий

Hardcoded URL

> Requires inference.sh CLI (`belt`). [Install instructions](https://raw.githubusercontent.com/infer

Line 15 links to raw GitHub installation instructions for the required CLI. This is an external network dependency and unpinned remote documentation risk.

Факторы риска

⚙️ Внешние команды (20)

🌐 Доступ к сети (2)