salesforce-developer

Fail

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: CRITICALEXTERNAL_DOWNLOADSMETADATA_POISONINGINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides a direct link to documentation hosted on an external domain (jeffallan.github.io) that has been explicitly flagged as malicious by automated URL scanners. Additionally, the CI/CD examples in references/deployment-devops.md include scripts that fetch and execute the Salesforce CLI binary from official Salesforce domains. While the Salesforce download is from a well-known service, the primary documentation link represents a high-risk external reference.
  • [METADATA_POISONING]: The skill metadata includes an author attribution and documentation link that point to a site identified as a malicious host. The SKILL.md file itself has been flagged by file reputation scanners as potentially malicious, likely due to the presence of these suspicious links.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data in the form of business requirements, data models, and user-provided code snippets to generate functional Salesforce logic.
  • Ingestion points: Salesforce requirement analysis and data model design steps in SKILL.md.
  • Boundary markers: No specific boundary markers or instructions are present to prevent the agent from obeying instructions embedded within processed business requirements.
  • Capability inventory: The skill is capable of generating code for network operations (Http.send), database modifications (DML), and automated deployment scripts using the Salesforce CLI.
  • Sanitization: The skill correctly advises the use of WITH SECURITY_ENFORCED and String.escapeSingleQuotes() when building dynamic queries, which mitigates some risks but does not eliminate the prompt injection surface.
  • [COMMAND_EXECUTION]: The skill provides detailed instructions for executing potentially sensitive shell commands via the Salesforce CLI, including authentication using JWT keys (sf org login jwt), metadata deployment (sf project deploy start), and data exports. While these are standard development practices, they involve high-privilege operations and the use of external credential files.
Recommendations
  • CRITICAL: 1 file(s) identified as malware by FileRep - DO NOT USE
  • AI detected serious security threats
  • Contains 2 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
Sep 5, 2026, 11:39 AM
Security Audit — agent-trust-hub — salesforce-developer