The Agent Skills Directory

[SAFE]: Analysis of the skill instructions and configuration confirms the absence of malicious patterns, obfuscation, or unauthorized data access.
[COMMAND_EXECUTION]: The skill requests access to the Bash tool in the YAML frontmatter. However, there are no instructions to execute shell commands or scripts. The tool access appears to be a standard configuration for the agent's environment.
[DATA_EXFILTRATION]: The skill logic involves gathering deal-related data (e.g., ACV, decision criteria) to assist in sales activities. There are no instructions to transmit this data to external servers or untrusted third parties.
[PROMPT_INJECTION]: The skill ingests external data regarding deal states. This provides a potential surface for indirect prompt injection, but the skill contains no patterns suggesting exploitation or bypass of safety filters.

deal-close