This report does not block or authorize the manifest or ZIP.
Static findings for external commands and network access are mostly true positives because the skill instructs users to install and run the inference.sh CLI. The remote shell installer is a confirmed high-risk supply-chain pattern, but I found no evidence of prompt injection or confirmed malicious intent in SKILL.md.
Report position
Historical report
Open audit history before using this report to install.
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 · 210 Lines analyzed
4 items shown for review
Limitations
This report does not claim runtime or sandbox execution and does not prove the absence of side effects.
01
Evidence chain
Follow the evidence from source binding to the install contract. Available evidence supports verification; it is not a safety guarantee.
TRUE_POSITIVE: The quick start instructs users to pipe a network download directly into sh. This can execute changed or compromised remote installer content before the user can inspect it.
The command pattern is explicit and semantically confirmed by the install instructions. The file later describes checksum verification, but the piped shell path still runs remote code directly.
TRUE_POSITIVE: The skill permits Bash execution for infsh and provides several infsh app run examples. This is core functionality, but it increases risk because model-assisted sessions can run external commands.
The external command surface is clearly present and intended. The examples use a named CLI rather than arbitrary shell construction, so the risk is elevated but not automatically malicious.
TRUE_POSITIVE: The skill sends React and Remotion source code, props, and render settings to inference.sh for processing. Users could expose proprietary video logic or sensitive data in props.
The network processing behavior is documented as the skill's purpose. It is legitimate for this skill, but the privacy and data exposure risk is real.
TRUE_POSITIVE: The related skills section tells users to run npx skills add commands for multiple external packages. This expands the supply-chain surface beyond the audited single skill file.
The commands are visible and relevant to package installation. They may be normal documentation, but they ask users to fetch additional unaudited content.
03
Risk findings
Confirmed security concerns are separated from items that still need review.
No confirmed security findings were recorded for this completed audit.
04
Expert evidence
Immutable subject identity, scanner metadata, dismissed matches, and source-level evidence.
Artifact subject
Marketplace commit
Unavailable
Content hash
Unavailable
Tree hash
Unavailable
Skill path
Unavailable
Audit payload hash
Unavailable
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.
FALSE_POSITIVE: The hardcoded Remotion and inference.sh URLs are documentation references. I did not find evidence that these links exfiltrate data by themselves.
The URLs appear in a documentation list and do not include encoded local data or secrets. The broader service dependency is covered separately.
FALSE_POSITIVE: The weak cryptography alert appears to match a checksum reference. SHA-256 checksum verification is not a weak algorithm indicator in this context.
Line 29 explicitly describes SHA-256 checksum verification for downloads. I did not find MD5, SHA-1, or custom cryptographic code in the reviewed file.