The Agent Skills Directory

[SAFE]: The skill uses standard Salesforce CLI (sf) commands to interact with the target environment for its intended purpose of configuring permission sets.
[SAFE]: File operations are restricted to relevant Salesforce metadata files (.permissionset-meta.xml, package.xml) and temporary directories created for inspection tasks.
[SAFE]: The script inspect-dataspace-scopes.sh uses standard Python libraries (xml.etree.ElementTree) to parse XML metadata locally. It does not execute remote code or perform network operations outside the Salesforce instance.
[SAFE]: Remote API interactions are directed to the official Salesforce Connect API endpoints (/services/data/v{apiVersion}/ssot/data-governance/...).
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided names for permission sets and dataspaces to generate configuration files. It uses specific XML and JSON templates, reducing the risk of accidental instruction execution. Severity is low as this is standard functionality for this type of skill.

platform-dataspace-access-configure