The Agent Skills Directory

[COMMAND_EXECUTION]: The skill automates tasks using the sf CLI and Node.js. It executes commands to query org metadata, verify package installations, and check connectivity status. These operations are scoped to the target Salesforce environment.
[DYNAMIC_EXECUTION]: The skill uses a template-based approach to generate anonymous Apex scripts for record creation and system updates. It employs a dedicated utility, scripts/render-apex.js, along with scripts/soql-escape.js to ensure that all user-provided data is properly escaped before being embedded in the executable code, effectively mitigating Apex and SOQL injection risks.
[INDIRECT_PROMPT_INJECTION]: The skill processes external inputs such as Sales Organization and KPI Set names. It manages the associated risks through a mandatory evidence chain:
Ingestion points: User-provided names are collected in Phase 6 and used in SOQL queries and Apex scripts.
Boundary markers: The skill uses single-quote delimiters in Apex variable declarations to isolate input data.
Capability inventory: The skill has the ability to write to the database and trigger external callouts via sf apex run and REST API calls.
Sanitization: All inputs are processed through escapeSoqlLiteral, which escapes backslashes and single quotes to prevent terminal character escapes.
[DATA_EXPOSURE]: Authentication is handled through the Salesforce CLI (sf). The scripts/sf-rest.js utility is designed to securely retrieve and use access tokens for API requests without surfacing them in logs or console output, following the principle of least exposure for credentials.

consumer-goods-rtr-datacloud-export-configure