The Agent Skills Directory

[SAFE]: The skill is purely instructional and provides a template for analyzing sales call data. It does not perform any high-risk operations such as network requests, shell command execution, or sensitive data access.
[DATA_EXPOSURE_&_EXFILTRATION]: The skill references a local file .agents/sales-context.md to provide context for sales analysis. This is a standard configuration file for the agent's domain and does not contain system-level secrets or credentials. No network operations were found that could exfiltrate this data.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input in the form of call transcripts and notes. While this is an ingestion surface for indirect prompt injection, the skill lacks high-privilege capabilities (like code execution or internet access) that could be exploited by such an injection. The impact is limited to the integrity of the generated analysis.
[REMOTE_CODE_EXECUTION]: There are no patterns of downloading or executing remote code. All logic is contained within the instructions provided to the AI agent.
[COMMAND_EXECUTION]: The skill does not use shell commands or dynamic context injection (!command) syntax.

call-debrief