The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the Salesforce CLI (sf data360) and a local utility script (node ../data360-orchestrate/scripts/diagnose-org.mjs). These operations are consistent with the skill's purpose of infrastructure management and use vendor-specific tooling.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection attack surface by processing metadata from external source systems.
Ingestion points: Metadata is retrieved via commands like sf data360 connection objects and sf data360 connection fields as specified in the SKILL.md workflow.
Boundary markers: There are no explicit instructions or delimiters provided to ensure the agent ignores potential instructions embedded in the external metadata.
Capability inventory: The agent has the capability to execute shell commands and interact with Salesforce orgs, which could be abused if malicious metadata is processed and followed as instructions.
Sanitization: The skill documentation does not outline any sanitization or validation steps for data retrieved from external connectors before it enters the agent's context.

data360-connect