The Agent Skills Directory

[SAFE]: The skill is entirely descriptive and instructional, containing no code, scripts, or automated command executions. All reference materials are standard markdown files documenting SaaS business strategies. All findings indicate the skill is used for domain-specific knowledge retrieval and strategic reasoning.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted inputs from sources such as project files and research notes. Ingestion points: Gtm Motion Design brief, decision audience, assumptions, and research notes as defined in SKILL.md. Boundary markers: The instructions explicitly command the agent to remain read-only and stop if required evidence is missing, which reduces the risk of malicious instructions being obeyed. Capability inventory: The skill has no executable capabilities, limited only to drafting and non-mutating checks. Sanitization: It directs the agent to separate facts from assumptions and produce a gap list, providing a structured approach that prevents the agent from treating untrusted data as verified evidence.\n- [EXTERNAL_DOWNLOADS]: The skill references informational articles from an external product design agency (Eleken) for practitioner cross-checks. These are static, plain-text links provided as supplementary context for the user and do not trigger automated code execution or downloads.

saas-gtm-motion-design