The Agent Skills Directory

[DATA_EXFILTRATION]: The skill accesses the user's configuration file located at the path defined by ${LOVSTUDIO_SKILLS_PROFILE:-$HOME/.lovstudio/skills/profile.json} to resolve settings such as workspace roots, output directories, and brand assets.
[COMMAND_EXECUTION]: The workflow requires invoking an external vendor tool, lovstudio-any2deck, to transform narrative outlines into rendered slide decks in PPTX and PDF formats.
[PROMPT_INJECTION]: The skill ingests untrusted content from user-provided outline.md files and evidence notes. This data is processed and passed to the rendering tool, which creates a surface for indirect prompt injection where malicious instructions hidden in the input files could attempt to manipulate the rendering outcome or workflow.
Ingestion points: User-provided outline.md narrative and evidence/source notes referenced in SKILL.md.
Boundary markers: The instructions do not define specific delimiters or security warnings to be used when interpolating the outline content into the rendering tool.
Capability inventory: The skill is capable of writing multiple files to the business-plan/ directory and executing the lovstudio-any2deck command as documented in the workflow.
Sanitization: The workflow instructions specify checks for narrative consistency and data accuracy but do not include explicit programmatic sanitization or filtering of the input content before processing.

lovstudio-bp-deck