prisma-7

Fail

Audited by Socket on Mar 1, 2026

3 alerts found:

AnomalySecurityObfuscated File
AnomalyLOW
800-guides/360-embed-studio-nextjs.md

This code is not obfuscated and contains no obvious malware constructs, but the provided API example is highly risky: it accepts arbitrary SQL queries from request bodies and executes them directly against the database, and the example sets CORS to '*' and does not perform authentication or authorization. In a real deployment this would allow anyone who can reach the endpoint to read, modify, or delete database data (data exfiltration and destructive operations). The tutorial correctly calls out the need to add auth and change CORS for production, but the example itself should not be used as-is in a public-facing environment. Recommend adding strict authentication/authorization, restricting CORS to trusted origins, validating or limiting allowed queries, and using a least-privilege database account for the endpoint.

Confidence: 90%Severity: 67%
SecurityMEDIUM
SKILL.md

The provided skill content is primarily documentation for a Prisma v7 integration and does not contain executable code, downloads, hardcoded credentials, or network endpoints. The main security concern is procedural: the manifest mandates spawning and running several named agents (transitive agent invocation), which creates a transitive trust and execution chain. If those agents are unreviewed or given broad permissions (file system access, environment variables, network), they could perform unintended actions or exfiltrate data. There is no explicit evidence of malware or obfuscated malicious code in the text, but the required multi-agent orchestration raises a moderate supply-chain/autonomy risk that should be controlled by reviewing the referenced agents and limiting their permissions before execution.

Confidence: 75%Severity: 75%
Obfuscated FileHIGH
references/pagination-offset.md

The analyzed code segment shows standard, well-structured offset-pagination implementations with additional pagination variants. There is no evidence of malicious behavior, backdoors, data exfiltration, or credential harvesting. The design includes input validation hints and defensive comments (e.g., max page size, cursor pagination for large offsets). Overall security risk is low with careful parameter handling; no malware indicators detected.

Confidence: 98%
Audit Metadata
Analyzed At
Mar 1, 2026, 09:32 PM
Package URL
pkg:socket/skills-sh/fusengine%2Fagents%2Fprisma-7%2F@a881fc9d825310eee4c2aa6c6f571838223ed840