The Agent Skills Directory

[SAFE]: No malicious behavior, obfuscation, or unauthorized access patterns were detected. The skill operates as a structured set of instructions and templates for content creation, delegating voice and brand checks to dedicated external skills.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests untrusted data in the form of 'Supporting material' (research, notes, drafts) from the user during the context gathering phase.
Ingestion points: SKILL.md Step 3 ('Supporting material', 'Topic').
Boundary markers: No specific delimiters or safety warnings for embedded instructions are defined in the workflow.
Capability inventory: The skill is limited to text drafting and the invocation of other specialized skills (creator-stack:voice, creator-stack:brand-guidelines). It does not demonstrate capabilities for code execution, filesystem modification, or network operations.
Sanitization: No explicit sanitization of user-provided context is performed before drafting. However, the structured multi-step workflow acts as a natural control, as the draft is subsequently processed by a 'Voice' skill and a 'Brand Guidelines' skill, which serves to align the output with specific predefined standards.

write