📦
Audit History
app-store-screenshots - 3 audits
Version comparison
Capability and finding changes across audited versions, newest first.
Audit version 3 Latest
Jul 12, 2026, 09:59 AM
The guide intentionally recommends external CLI installation and remote inference commands, creating supply-chain and third-party processing risks. Most detections are Markdown fences or prose references without execution behavior.
1
Files scanned
271
Lines analyzed
13
Review items
0
False positives ignored
Confirmed security concerns (1)
Medium
External Processing Without Privacy Guidance
The workflow sends prompts and potentially local app screenshots to external inference models without warning users to remove sensitive or confidential content.
The guide names an external inference service and supplies local image filenames to a remote app command. No privacy or data-handling warning appears.
Capability review items (10)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Medium
Ruby/shell backtick execution
> **Install the belt CLI skill:** `npx skills add belt-sh/cli`
Line 7 explicitly instructs users to execute npx against a third-party package source. This creates external command and package supply-chain exposure.
Medium
SKILL.md:17-26SKILL.md:137-144SKILL.md:148-154SKILL.md:158-165SKILL.md:169-175SKILL.md:206-211SKILL.md:262-267
Ruby/shell backtick execution · 7 occurrences
```bash
The cited block contains belt login and belt app run commands. These invoke an installed external CLI and submit an inference job.
Medium
Ruby/shell backtick execution
```
Although line 267 is a closing fence, the cited range also tells users to run belt app store. This is an explicit external CLI command.
Low
Hardcoded URL
Create app store screenshots and preview videos via [inference.sh](https://inference.sh) CLI.
The text states that generation occurs through inference.sh, and later commands submit jobs to that service. External processing is an intended network dependency.
Risk Factors
⚙️ External commands (20)
SKILL.md:7 SKILL.md:15 SKILL.md:17-26 SKILL.md:26-83 SKILL.md:83-97 SKILL.md:97-121 SKILL.md:121-131 SKILL.md:131-137 SKILL.md:137-144 SKILL.md:144-148 SKILL.md:148-154 SKILL.md:154-158 SKILL.md:158-165 SKILL.md:165-169 SKILL.md:169-175 SKILL.md:175-206 SKILL.md:206-211 SKILL.md:211-262 SKILL.md:262-267 SKILL.md:267-269
🌐 Network access (2)
Audited by: codex
Jul 12, 2026, 09:59 AM
The guide intentionally recommends external CLI installation and remote inference commands, creating supply-chain and third-party processing risks. Most detections are Markdown fences or prose references without execution behavior.
1
Files scanned
271
Lines analyzed
13
Review items
0
False positives ignored
Confirmed security concerns (1)
Medium
External Processing Without Privacy Guidance
The workflow sends prompts and potentially local app screenshots to external inference models without warning users to remove sensitive or confidential content.
The guide names an external inference service and supplies local image filenames to a remote app command. No privacy or data-handling warning appears.
Capability review items (10)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Medium
Ruby/shell backtick execution
> **Install the belt CLI skill:** `npx skills add belt-sh/cli`
Line 7 explicitly instructs users to execute npx against a third-party package source. This creates external command and package supply-chain exposure.
Medium
SKILL.md:17-26SKILL.md:137-144SKILL.md:148-154SKILL.md:158-165SKILL.md:169-175SKILL.md:206-211SKILL.md:262-267
Ruby/shell backtick execution · 7 occurrences
```bash
The cited block contains belt login and belt app run commands. These invoke an installed external CLI and submit an inference job.
Medium
Ruby/shell backtick execution
```
Although line 267 is a closing fence, the cited range also tells users to run belt app store. This is an explicit external CLI command.
Low
Hardcoded URL
Create app store screenshots and preview videos via [inference.sh](https://inference.sh) CLI.
The text states that generation occurs through inference.sh, and later commands submit jobs to that service. External processing is an intended network dependency.
Risk Factors
⚙️ External commands (20)
SKILL.md:7 SKILL.md:15 SKILL.md:17-26 SKILL.md:26-83 SKILL.md:83-97 SKILL.md:97-121 SKILL.md:121-131 SKILL.md:131-137 SKILL.md:137-144 SKILL.md:144-148 SKILL.md:148-154 SKILL.md:154-158 SKILL.md:158-165 SKILL.md:165-169 SKILL.md:169-175 SKILL.md:175-206 SKILL.md:206-211 SKILL.md:211-262 SKILL.md:262-267 SKILL.md:267-269
🌐 Network access (2)
Audited by: codex
Jul 10, 2026, 11:11 AM
Most shell-execution alerts are Markdown delimiters or text examples, but nine entries contain genuine CLI commands for installation, generation, or discovery. No malicious or prompt-injection intent was found, although remote processing may expose prompts and app screenshots to a third-party service.
1
Files scanned
271
Lines analyzed
12
Review items
0
False positives ignored
Confirmed security concerns (1)
Medium
Potential Third-Party Upload of App Screenshots
The before-and-after workflow passes local screenshot filenames to an external inference application, which may expose unreleased interfaces or user data.
The command invokes an external application and names local image files as inputs. Upload is strongly implied, although transport details are not shown.
Capability review items (9)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Medium
Ruby/shell backtick execution
> **Install the belt CLI skill:** `npx skills add belt-sh/cli`
Line 7 directs users to execute an npx command that installs a remote skill. This is genuine external command execution with dependency supply-chain exposure.
Medium
SKILL.md:17-26SKILL.md:137-144SKILL.md:148-154SKILL.md:158-165SKILL.md:169-175SKILL.md:206-211SKILL.md:262-267
Ruby/shell backtick execution · 7 occurrences
```bash
The fenced block runs belt login and a belt app command against an external inference service. The commands are intentional, but they execute external software and transmit input.
Medium
Ruby/shell backtick execution
```
The cited range includes the inline belt app store command. It is a real external CLI invocation, although its apparent purpose is limited to listing available applications.
Risk Factors
⚙️ External commands (20)
SKILL.md:7 SKILL.md:15 SKILL.md:17-26 SKILL.md:26-83 SKILL.md:83-97 SKILL.md:97-121 SKILL.md:121-131 SKILL.md:131-137 SKILL.md:137-144 SKILL.md:144-148 SKILL.md:148-154 SKILL.md:154-158 SKILL.md:158-165 SKILL.md:165-169 SKILL.md:169-175 SKILL.md:175-206 SKILL.md:206-211 SKILL.md:211-262 SKILL.md:262-267 SKILL.md:267-269
🌐 Network access (2)
Audited by: codex