The static findings are false positives caused by Markdown inline code, fenced command examples, documentation links, and a documented token-storage path. No prompt injection, hidden exfiltration intent, or undisclosed filesystem access was found in SKILL.md; the skill openly relies on the RunComfy CLI and external RunComfy API.
The static findings are false positives caused by Markdown inline code, fenced command examples, documentation links, and a documented token-storage path. No prompt injection, hidden exfiltration intent, or undisclosed filesystem access was found in SKILL.md; the skill openly relies on the RunComfy CLI and external RunComfy API.
Static analysis flagged command, URL, filesystem, and weak-cryptography patterns in SKILL.md. Most matches are markdown examples or substring false positives. Real risk comes from scoped RunComfy CLI execution, hosted API calls, and local token storage.
1
Files scanned
189
Lines analyzed
7
Review items
1
False positives ignored
Capability review items (4)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Shell Metacharacter Mentions Are Defensive Documentation
The command-substitution alert appears in a security note about shell metacharacters. It is not an executable user-controlled command.
The flagged syntax is explanatory prose about shell injection. No command at this line expands user prompt content.
Static false positives ignored (1)
These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.
Static analysis flagged 93 potential issues, but evaluation reveals all external_commands patterns are markdown code formatting (backticks for inline code), weak_crypto detections are numeric strings (music durations/frequencies not crypto), and network URLs are legitimate RunComfy documentation links. The skill invokes the runcomfy CLI to call the ElevenLabs Music API - a standard music generation wrapper with no malicious behavior detected. All shell commands are pre-defined CLI invocations; no user input injection vectors exist.
Numeric Values Misidentified as Weak Cryptographic Algorithms
Static analysis flagged 10 instances of 'weak cryptographic algorithm'. Evaluation reveals these are numeric strings representing music duration, frequency, and bar counts (e.g., '44.1 kHz', '5 seconds', '8 bars', '120 BPM'). No cryptographic algorithms are used in this skill.
Pattern matching flagged numeric strings as crypto algorithms. No crypto usage exists in this skill - it only calls the runcomfy CLI to invoke the ElevenLabs Music API.
Capability review items (1)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
The skill documents storing API tokens in ~/.config/runcomfy/token.json. This is standard CLI token storage behavior, documented with appropriate file permissions (mode 0600).
Standard token storage pattern for CLI tools. File permissions are properly documented, and the skill does not expose the token value.
Static false positives ignored (2)
These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.
Markdown Backticks Misidentified as Shell Commands
Static analysis flagged 69 instances of 'Ruby/shell backtick execution'. Evaluation shows these are Markdown code formatting syntax used throughout documentation, not executable shell commands. The backticks surround CLI examples like `runcomfy run elevenlabs/...` for documentation display.
All backtick instances in SKILL.md are clearly markdown inline code formatting for documenting CLI commands. No actual shell execution occurs from these patterns.
The skill contains hardcoded URLs to RunComfy documentation pages. These are necessary documentation links for a CLI wrapper skill. URLs target legitimate runcomfy.com domains.
URLs are legitimate RunComfy documentation links. No third-party or untrusted endpoints present.