📦
Audit History
image-to-video - 3 audits
Version comparison
Capability and finding changes across audited versions, newest first.
Audit version 3 Latest
Jul 24, 2026, 01:35 AM
Most static alerts are false positives caused by Markdown backticks, placeholder URLs, and standard configuration paths. Eight findings confirm intended installation or execution of external CLI commands. The skill also sends prompts and user media references to a third-party model service.
1
Files scanned
206
Lines analyzed
12
Review items
0
False positives ignored
Confirmed security concerns (1)
Medium
User Media Sent to a Third-Party Service
The CLI submits prompts to RunComfy, and its model server fetches user-provided image, audio, and video URLs. This creates a disclosed privacy and authorization risk.
The skill explicitly describes API submission, remote media fetching, polling, and output downloads. The third-party processing path is direct and unambiguous.
Capability review items (8)
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
model. Calls `runcomfy run <vendor>/<model>/image-to-video` (or
The skill explicitly states that it calls the local RunComfy CLI. This is intended external process execution with model identifiers selected by the workflow.
Medium
Ruby/shell backtick execution · 4 occurrences
```bash
The fenced example invokes `npx skills add` against a community repository. Executing it downloads and installs remote skill content.
Medium
Ruby/shell backtick execution
1. **RunComfy CLI** — `npm i -g @runcomfy/cli`
The prerequisite directs users to install `@runcomfy/cli` globally with npm. This executes package installation code from an external registry.
Medium
Ruby/shell backtick execution
2. **RunComfy account** — `runcomfy login` opens a browser device-code flow.
The prerequisite directs users to run `runcomfy login`, which starts an external CLI and browser authentication flow.
Medium
Ruby/shell backtick execution
The skill picks one of HappyHorse 1.0 I2V / Wan 2.7 t2v+audio / Seedance 2.0 Pro based on user inten
The line states that the workflow invokes `runcomfy run`, polls a remote request, and downloads results. This confirms intended external process and network activity.
Risk Factors
⚙️ External commands (50)
SKILL.md:9 SKILL.md:13 SKILL.md:27-29 SKILL.md:29-38 SKILL.md:38-39 SKILL.md:39 SKILL.md:48-49 SKILL.md:49-50 SKILL.md:50-57 SKILL.md:57-63 SKILL.md:63-64 SKILL.md:64-65 SKILL.md:65 SKILL.md:66-67 SKILL.md:67-68 SKILL.md:68-74 SKILL.md:74-81 SKILL.md:81-93 SKILL.md:93-95 SKILL.md:95 SKILL.md:103-104 SKILL.md:104-105 SKILL.md:105 SKILL.md:106 SKILL.md:107 SKILL.md:108-109 SKILL.md:109-113 SKILL.md:113-122 SKILL.md:122-127 SKILL.md:127-128 SKILL.md:128 SKILL.md:129-135 SKILL.md:135-143 SKILL.md:143-144 SKILL.md:144 SKILL.md:145 SKILL.md:146 SKILL.md:147 SKILL.md:148-149 SKILL.md:149 SKILL.md:150-151 SKILL.md:151-155 SKILL.md:155-165 SKILL.md:165-169 SKILL.md:169 SKILL.md:170-180 SKILL.md:180 SKILL.md:197 SKILL.md:201 SKILL.md:202-204
🌐 Network access (8)
📁 Filesystem access (7)
Audited by: codex
Jul 15, 2026, 10:05 AM
Most static findings are lexical false positives caused by Markdown backticks, URL placeholders, and documented configuration paths. The skill intentionally runs RunComfy CLI commands and installs external tooling, so those command findings remain confirmed at medium severity. It also sends prompts and user-supplied media URLs to RunComfy for remote processing.
1
Files scanned
206
Lines analyzed
13
Review items
0
False positives ignored
Confirmed security concerns (1)
Medium
User Media Sent to Third-Party Service
The workflow sends prompts and media URLs to RunComfy, which fetches referenced content and returns downloadable outputs. Sensitive assets leave the local environment.
The source explicitly says the CLI posts requests, RunComfy servers fetch referenced media, and generated files are downloaded from RunComfy domains.
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
model. Calls `runcomfy run <vendor>/<model>/image-to-video` (or
The text explicitly states that the skill invokes the local RunComfy CLI. This is intended external process execution with network and filesystem effects.
Medium
Ruby/shell backtick execution · 4 occurrences
```bash
The fenced command runs npx to install a GitHub-hosted skill globally. This executes third-party tooling and changes global skill configuration.
Medium
Ruby/shell backtick execution
1. **RunComfy CLI** — `npm i -g @runcomfy/cli`
This instruction globally installs the RunComfy CLI from npm. Package installation executes third-party lifecycle code and changes the host environment.
Medium
Ruby/shell backtick execution
2. **RunComfy account** — `runcomfy login` opens a browser device-code flow.
This instruction executes runcomfy login and begins an authentication flow. The command also creates credential state through the external CLI.
Medium
Ruby/shell backtick execution
The skill picks one of HappyHorse 1.0 I2V / Wan 2.7 t2v+audio / Seedance 2.0 Pro based on user inten
The workflow explicitly invokes runcomfy, polls a remote API, and downloads results. This confirms intended external command execution and side effects.
Medium
Ruby/shell backtick execution
- **Token storage**: `runcomfy login` writes the API token to `~/.config/runcomfy/token.json` with m
The documented runcomfy login command creates token storage under the user profile. It is intended external execution that changes credential state.
Risk Factors
⚙️ External commands (50)
SKILL.md:9 SKILL.md:13 SKILL.md:27-29 SKILL.md:29-38 SKILL.md:38-39 SKILL.md:39 SKILL.md:48-49 SKILL.md:49-50 SKILL.md:50-57 SKILL.md:57-63 SKILL.md:63-64 SKILL.md:64-65 SKILL.md:65 SKILL.md:66-67 SKILL.md:67-68 SKILL.md:68-74 SKILL.md:74-81 SKILL.md:81-93 SKILL.md:93-95 SKILL.md:95 SKILL.md:103-104 SKILL.md:104-105 SKILL.md:105 SKILL.md:106 SKILL.md:107 SKILL.md:108-109 SKILL.md:109-113 SKILL.md:113-122 SKILL.md:122-127 SKILL.md:127-128 SKILL.md:128 SKILL.md:129-135 SKILL.md:135-143 SKILL.md:143-144 SKILL.md:144 SKILL.md:145 SKILL.md:146 SKILL.md:147 SKILL.md:148-149 SKILL.md:149 SKILL.md:150-151 SKILL.md:151-155 SKILL.md:155-165 SKILL.md:165-169 SKILL.md:169 SKILL.md:170-180 SKILL.md:180 SKILL.md:197 SKILL.md:201 SKILL.md:202-204
🌐 Network access (8)
📁 Filesystem access (7)
Audited by: codex
Jul 15, 2026, 10:05 AM
Most static findings are lexical false positives caused by Markdown backticks, URL placeholders, and documented configuration paths. The skill intentionally runs RunComfy CLI commands and installs external tooling, so those command findings remain confirmed at medium severity. It also sends prompts and user-supplied media URLs to RunComfy for remote processing.
1
Files scanned
206
Lines analyzed
13
Review items
0
False positives ignored
Confirmed security concerns (1)
Medium
User Media Sent to Third-Party Service
The workflow sends prompts and media URLs to RunComfy, which fetches referenced content and returns downloadable outputs. Sensitive assets leave the local environment.
The source explicitly says the CLI posts requests, RunComfy servers fetch referenced media, and generated files are downloaded from RunComfy domains.
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
model. Calls `runcomfy run <vendor>/<model>/image-to-video` (or
The text explicitly states that the skill invokes the local RunComfy CLI. This is intended external process execution with network and filesystem effects.
Medium
Ruby/shell backtick execution · 4 occurrences
```bash
The fenced command runs npx to install a GitHub-hosted skill globally. This executes third-party tooling and changes global skill configuration.
Medium
Ruby/shell backtick execution
1. **RunComfy CLI** — `npm i -g @runcomfy/cli`
This instruction globally installs the RunComfy CLI from npm. Package installation executes third-party lifecycle code and changes the host environment.
Medium
Ruby/shell backtick execution
2. **RunComfy account** — `runcomfy login` opens a browser device-code flow.
This instruction executes runcomfy login and begins an authentication flow. The command also creates credential state through the external CLI.
Medium
Ruby/shell backtick execution
The skill picks one of HappyHorse 1.0 I2V / Wan 2.7 t2v+audio / Seedance 2.0 Pro based on user inten
The workflow explicitly invokes runcomfy, polls a remote API, and downloads results. This confirms intended external command execution and side effects.
Medium
Ruby/shell backtick execution
- **Token storage**: `runcomfy login` writes the API token to `~/.config/runcomfy/token.json` with m
The documented runcomfy login command creates token storage under the user profile. It is intended external execution that changes credential state.
Risk Factors
⚙️ External commands (50)
SKILL.md:9 SKILL.md:13 SKILL.md:27-29 SKILL.md:29-38 SKILL.md:38-39 SKILL.md:39 SKILL.md:48-49 SKILL.md:49-50 SKILL.md:50-57 SKILL.md:57-63 SKILL.md:63-64 SKILL.md:64-65 SKILL.md:65 SKILL.md:66-67 SKILL.md:67-68 SKILL.md:68-74 SKILL.md:74-81 SKILL.md:81-93 SKILL.md:93-95 SKILL.md:95 SKILL.md:103-104 SKILL.md:104-105 SKILL.md:105 SKILL.md:106 SKILL.md:107 SKILL.md:108-109 SKILL.md:109-113 SKILL.md:113-122 SKILL.md:122-127 SKILL.md:127-128 SKILL.md:128 SKILL.md:129-135 SKILL.md:135-143 SKILL.md:143-144 SKILL.md:144 SKILL.md:145 SKILL.md:146 SKILL.md:147 SKILL.md:148-149 SKILL.md:149 SKILL.md:150-151 SKILL.md:151-155 SKILL.md:155-165 SKILL.md:165-169 SKILL.md:169 SKILL.md:170-180 SKILL.md:180 SKILL.md:197 SKILL.md:201 SKILL.md:202-204
🌐 Network access (8)
📁 Filesystem access (7)
Audited by: codex