The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses clear, task-oriented instructions focused on sales methodology. No attempts to bypass safety filters, extract system prompts, or override agent constraints were found.
[DATA_EXFILTRATION]: No network operations, hardcoded credentials, or sensitive file path accesses were detected. The skill operates entirely within the chat context to produce text-based deliverables.
[OBFUSCATION]: All files consist of plain-text Markdown and JSON. No Base64 encoding, zero-width characters, homoglyphs, or other hidden content techniques were identified.
[REMOTE_CODE_EXECUTION]: The skill does not include any scripts (Python, Node.js, Shell) or instructions to download and execute external code. It is a 'no-code' instructional skill.
[COMMAND_EXECUTION]: No shell commands, privilege escalation attempts (sudo), or system-level modifications are present in the instruction set or metadata.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user-provided sales context (product info, deal history), it lacks the dangerous capabilities (network, file-system writes, code execution) required to exploit this ingestion surface. It includes instructions for the agent to ask clarifying questions rather than blindly following embedded data directives.

sales-qualification