The Agent Skills Directory

[SAFE]: The skill instructions are focused on creative writing tasks, specifically creating and updating character profiles in a characters/ directory. No malicious commands or unsafe data handling were detected.
[INDIRECT_PROMPT_INJECTION]: The skill reads user-provided content from story.md and existing character files to maintain consistency. While these files are untrusted inputs, the skill's limited capabilities (writing text to markdown files) and the use of a predefined YAML template minimize the risk of behavioral override.
Ingestion points: story.md, characters/_index.md, and files in characters/.
Boundary markers: The skill follows a strict template in references/character-template.md to format character files.
Capability inventory: Reading and writing local markdown files within the story project structure.
Sanitization: None explicitly defined, but the workflow relies on structured YAML frontmatter for data consistency.

character-management