The Agent Skills Directory

[PROMPT_INJECTION]: No patterns of bypass attempts, role-play injections, or instructions to ignore safety guidelines were detected. The instructions are focused entirely on the business logic of scheduling reviews.
[DATA_EXFILTRATION]: While the skill accesses sensitive CRM data (deals, contact information, activity history), it does so using platform-native execute_action tools. There is no evidence of data being sent to external or untrusted third-party domains.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: Data enters the agent context through CRM tools like get_deal, get_contact, and get_deal_activities (SKILL.md).
Boundary markers: The skill lacks explicit delimiters or instructions to ignore potential commands embedded within the CRM data.
Capability inventory: The skill uses crm, email, and calendar capabilities to generate structured email drafts and meeting metadata. It does not possess capabilities for file writing or system command execution.
Sanitization: There is no evidence of sanitization or filtering of the CRM data before it is interpolated into the email templates.
[REMOTE_CODE_EXECUTION]: The skill does not download external scripts, install packages, or use dynamic execution functions like eval() or exec().
[COMMAND_EXECUTION]: No shell commands, subprocess spawning, or system-level operations are present in the skill definition.

QBR Scheduler