Versioned security assessment

Report ID: SA-3E4B6C31

7/9/2026, 7:58:59 AM

happyhorse-1-0 security assessment v9

Skill Security Certification Report

Audit History

Audit model: codex Latest published report

Skill name: happyhorse-1-0
Version: v9
Maintainer: agentspace-so
Source: github.com/agentspace-so/runcomfy-agent-skills/tree/main/happyhorse-1-0/
Coverage: 1 Files scanned · 196 Lines analyzed
Policy version: Unavailable

Highest confirmed finding severity

Medium

1 confirmed security finding requires attention.

Installation context

Check the current Skill page

This page summarizes report evidence only. The Skill page provides the canonical install advisory.

Open current Skill page

This report does not block or authorize the manifest or ZIP.

The skill has expected external-command and network behavior because it runs the RunComfy CLI to submit video jobs. Most static hits are false positives from Markdown backticks and ellipsis in documentation, but CLI execution and bearer-token API submission remain confirmed risks. No evidence found of prompt injection text in the skill file.

Report position

Latest published report

Latest refers to the report sequence, not to artifact currentness.

Audit attestation

Not attestable

The required immutable binding is incomplete.

Human verification

Not verified

No human verification is recorded for this report.

Coverage

1 Files scanned · 196 Lines analyzed

11 items shown for review

Limitations

This report does not claim runtime or sandbox execution and does not prove the absence of side effects.

Evidence chain

Follow the evidence from source binding to the install contract. Available evidence supports verification; it is not a safety guarantee.

Source

Commit and path bound
Artifact

Content and tree hashes bound
Audit

Complete
Install contract

Open manifest to verify
Open manifest

Capabilities observed

Observed means this report recorded supporting evidence. Not recorded does not prove that a capability is absent.

Contains scripts

May execute code included with the Skill.

Not recorded by this audit

Network access

May connect to external services.

Observed in 4 evidence locations

Filesystem access

May read or write local files.

Observed in 4 evidence locations

Env variables

May read values from the process environment.

Not recorded by this audit

External commands

May invoke commands or programs outside the Skill.

Observed in 40 evidence locations

Capability review items (10)

Medium

SKILL.md:10

Ruby/shell backtick execution

Calls `runcomfy run happyhorse/happyhorse-1-0/text-to-video` through

The skill explicitly says it calls the local RunComfy CLI. This is expected behavior, but it is still local external command execution against a third-party service.

Medium

SKILL.md:24-26

Ruby/shell backtick execution

```bash

The fenced setup example includes an npx install command that installs a skill globally from a GitHub source. Running installer commands from community documentation is a real external command risk.

Medium

SKILL.md:44-45

Ruby/shell backtick execution

1. **RunComfy CLI** — `npm i -g @runcomfy/cli`

The prerequisite instructs users to run npm to install the RunComfy CLI globally. This is expected setup, but it executes package-manager code on the local machine.

Medium

SKILL.md:45-46

Ruby/shell backtick execution

2. **RunComfy account** — `runcomfy login` opens a browser device-code flow.

The prerequisite instructs users to run the RunComfy login command. The command initiates authentication and can affect local credential state.

Medium

SKILL.md:65-69

Ruby/shell backtick execution

```bash

The example runs the RunComfy CLI with user-supplied prompt content and writes outputs to a local directory. Arguments are structured, but the command still executes locally and contacts a third-party API.

Medium

SKILL.md:73-82

Ruby/shell backtick execution

```bash

The vertical video example runs the RunComfy CLI and disables the watermark option. It is a legitimate workflow, but it remains local command execution with network submission.

Medium

SKILL.md:86-90

Ruby/shell backtick execution

```bash

The cheaper test example runs the RunComfy CLI with prompt input and output directory arguments. This confirms expected local command execution rather than a pure documentation-only skill.

Medium

SKILL.md:178-179

Ruby/shell backtick execution

1. The skill invokes `runcomfy run happyhorse/happyhorse-1-0/text-to-video` with a JSON body matchin

The implementation notes state that the skill invokes the RunComfy CLI with a JSON body. This directly confirms the external command behavior described by the skill.

Medium

SKILL.md:191

Ruby/shell backtick execution

- **Token storage**: `runcomfy login` writes the API token to `~/.config/runcomfy/token.json` with m

The security section references runcomfy login as the command that writes the API token. The command is legitimate setup, but it changes local credential state and should be treated as a real command risk.

Medium

SKILL.md:179

Hardcoded URL

2. The CLI POSTs to `https://model-api.runcomfy.net/v1/models/happyhorse/happyhorse-1-0/text-to-vide

The skill documents a hardcoded RunComfy Model API endpoint that receives the user bearer token. The endpoint is expected for this skill, but it is a real third-party network and credential transmission risk.

Risk findings

Confirmed security concerns are separated from items that still need review.

Confirmed security concerns (1)

RISK-001 Medium

SKILL.md:179 SKILL.md:192

Third-Party Prompt and Token Transmission

The skill documents that the CLI posts to the RunComfy Model API with the user bearer token and sends the prompt through the CLI input boundary. This is required for the workflow, but users should treat prompts and credentials as third-party data exposure.

The documentation explicitly states that requests use the user bearer token and that user prompts are passed through the CLI to the Model API. The behavior is transparent, but the data-flow risk is real.

Remediation

Suggested fixes recorded by this audit. Applying them is the maintainer’s responsibility.

FIX-001
Medium
External CLI execution and installer commands

Require explicit user confirmation before installing or running npx, npm, or runcomfy commands, and display the exact arguments before execution.
FIX-002
Medium
Third-party prompt and credential transmission

Clearly disclose that prompts and bearer tokens are sent to RunComfy, recommend scoped tokens, and warn users not to submit sensitive prompts.
FIX-003
Low
Local token storage

Document the token file path, expected permissions, token revocation steps, and the RUNCOMFY_TOKEN alternative for CI environments.
FIX-004
Low
Generated output downloads

Keep the host allowlist and file-size cap, sanitize output filenames, and constrain downloads to the requested output directory.

Expert evidence

Immutable subject identity, scanner metadata, dismissed matches, and source-level evidence.

Artifact subject

Marketplace commit: 3e4b6c31a74a3bd1a291c98cf585d720cb9fbc88
Content hash: eaca8bb73e047efd0d2ab742a82dad5d52a0bda073e4fb301b3a651046475e00
Tree hash: 72563048675db8e29704ddaeabdeea5f8a387d4fa452f6936c2b908332df09e5
Skill path: skills/agentspace-so/happyhorse-1-0
Audit payload hash: 20fb63a4d72c4a4fb3f2e7c4ae662fdb

Analysis metadata

Audit model: codex

Analysis state: Complete

Scope is limited to the recorded files, lines, methods, and evidence. No runtime or sandbox execution is claimed.

Verify and export

JSON report CITATION.cff SARIF 2.1.0 Signed installation manifest Source

The manifest and lockfile bind install artifacts to cryptographic hashes. This integrity claim is separate from the security assessment.

Audit attestation: not_attestable