platform-agentexchange-partner-offers-configure

Pass

Audited by Gen Agent Trust Hub on Aug 7, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses standard CLI tools (sf and jq) to perform routine Salesforce administrative tasks, such as querying org data and deploying metadata files. These operations are well-scoped to the skill's stated purpose.
  • [DATA_EXFILTRATION]: Analysis found no evidence of sensitive data being sent to external or unauthorized domains. All data operations are confined to the user's authenticated Salesforce environment via the CLI.
  • [PROMPT_INJECTION]: The instructions do not contain any patterns intended to bypass safety protocols, override agent behavior, or extract system prompts. The language is strictly instructional and task-oriented.
  • [REMOTE_CODE_EXECUTION]: The skill does not download or execute scripts from remote sources. It relies entirely on pre-installed system tools and local project files.
  • [CREDENTIALS_UNSAFE]: No hardcoded credentials or secrets were found. The skill correctly utilizes the existing authentication context of the Salesforce CLI.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 7, 2026, 01:00 PM
Security Audit — agent-trust-hub — platform-agentexchange-partner-offers-configure