The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the shell and the Salesforce CLI (sf) to perform project management tasks, including retrieving metadata from orgs, deploying Apex code, and executing queries. These operations are the primary purpose of the skill and are performed using official vendor tools.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks because its core function involves analyzing production conversation logs which contain untrusted user messages.
Ingestion points: Conversation data is ingested from Data Cloud Data Model Objects (DMOs) such as ssot__AiAgentInteractionMessage__dlm and ssot__AiAgentMoment__dlm (detailed in references/stdm-schema.md).
Boundary markers: The instructions for reconstructing and analyzing conversations (found in references/stdm-queries.md) do not explicitly mandate the use of delimiters or 'ignore' instructions for the log content.
Capability inventory: The agent has the ability to execute shell commands via Bash and modify project files using Edit and Write tools.
Sanitization: The skill does not provide specific procedures for sanitizing or escaping retrieved user utterances before the agent evaluates them.

observing-agentforce