The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts check_perms.py and setup_perms.py execute shell commands using the sf CLI to interact with Salesforce organizations. This includes querying record data and deploying metadata. This behavior is the primary intended function of the skill for performing org audits and provisioning.
[DYNAMIC_EXECUTION]: The setup_perms.py script generates Salesforce permission set XML metadata at runtime in a temporary directory and uses the sf CLI to deploy it to the target organization. This dynamic creation of configuration files is necessary for its advertised functionality.
[INDIRECT_PROMPT_INJECTION]: The skill ingests metadata and record information from Salesforce organizations via the sf CLI. While this data is external, it is handled as structured JSON data for the purpose of generating a permission scorecard and does not appear to be used in a way that would influence agent behavior.
Ingestion points: User record data and permission set metadata from Salesforce organizations via sf data query in check_perms.py and setup_perms.py.
Boundary markers: Not applicable; data is processed as structured JSON within the Python scripts rather than being directly interpolated into an LLM prompt.
Capability inventory: The skill has the capability to execute shell commands and modify Salesforce organization metadata via the sf CLI.
Sanitization: Input parameters for CLI commands are passed as lists to subprocess.run, which prevents shell command injection.

sra-perm-check