The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Salesforce CLI (sf) to execute queries, retrieve metadata, and deploy Apex code to the user's organization. These operations utilize official vendor tools and are performed within the context of the user's authenticated environment.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it analyzes conversation logs and LLM prompts retrieved from Salesforce Data Cloud. 1. Ingestion points: Interaction data and LLM prompts are retrieved via the AgentforceOptimizeService Apex class from interaction DMOs. 2. Boundary markers (absent): The skill does not use specific delimiters or instructions to the agent to treat the retrieved log data as untrusted or non-executable content. 3. Capability inventory: The agent has access to shell tools (Bash), file manipulation (Edit, Write), and Salesforce deployment capabilities. 4. Sanitization (absent): No sanitization is applied to the retrieved message content before analysis.
[EXTERNAL_DOWNLOADS]: The skill retrieves agent configuration bundles and session telemetry from the connected Salesforce cloud environment. These downloads are performed from trusted vendor-managed sources.

agentforce-observe