The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data such as concepts, math expressions, algorithm descriptions, and research paper content to generate animation scripts.
Ingestion points: The skill instructions in SKILL.md explicitly direct the agent to take "Topic/concept", "Math expressions", "Algorithm description", and "Research paper" content as inputs for its various modes.
Boundary markers: The skill lacks explicit boundary markers or "ignore previous instructions" warnings when interpolating these user-provided inputs into the generated Python code.
Capability inventory: The skill grants the agent the ability to write arbitrary Python code to a file (script.py) and execute it via the manim CLI, which can also trigger LaTeX (pdflatex) rendering, and perform video manipulation via ffmpeg.
Sanitization: There are no instructions for sanitizing, escaping, or validating external content before it is placed into the executable script context.
[DYNAMIC_EXECUTION]: The skill is designed to generate Python code at runtime and execute it to produce video output.
Evidence: SKILL.md describes a production pipeline where the agent writes a script.py file containing Manim scene classes and executes it using commands like manim -ql script.py SceneName.
Context: This dynamic execution of generated code is the primary intended purpose of the skill and is required for its function as a video production pipeline.

manim-video