The Agent Skills Directory

[SAFE]: The skill implements a structured methodology for business acquisitions and readiness assessments, providing checklists for financial, technical, and legal due diligence.
[COMMAND_EXECUTION]: The multi-agent section utilizes a Task orchestration syntax to delegate research activities. These tasks are scoped to information gathering from public sources and do not allow for arbitrary shell command execution.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it interpolates user-provided data into sub-agent prompts.
Ingestion points: {target_company}, {category}, and {buyer_type} placeholders in SKILL.md are populated by user input.
Boundary markers: None identified; inputs are interpolated directly into the research prompt strings.
Capability inventory: The sub-agents perform external research and planning tasks, but do not have file-write or system-level capabilities.
Sanitization: No explicit validation or sanitization of input placeholders is present in the prompt templates.
[DATA_EXFILTRATION]: No unauthorized data access or external transmission of sensitive information was detected. The research activities target public profiles and platforms as intended by the skill's business purpose.

ma-playbook