The Agent Skills Directory

[PROMPT_INJECTION]: The skill's instructions in SKILL.md and individual rules files focus exclusively on teaching the agent how to use the Manim Community Edition. No override markers, role-play injections, or attempts to bypass safety filters were found.
[DATA_EXFILTRATION]: No hardcoded credentials, sensitive file paths, or unauthorized network operations are present. The skill correctly identifies and recommends standard practices like using configuration files (manim.cfg) without exposing private data.
[OBFUSCATION]: The skill uses clear, human-readable Markdown and Python code. No Base64, hex encoding, zero-width characters, or homoglyph attacks were detected.
[EXTERNAL_DOWNLOADS]: The skill references legitimate, well-known software including the 'manim' library, 'numpy', and 'scipy'. These are standard requirements for the animation engine described and originate from trusted registries.
[REMOTE_CODE_EXECUTION]: No patterns of downloading and executing remote scripts (e.g., curl|bash) were found. The CLI instructions provided in rules/cli.md are standard operational commands for the Manim tool.
[DYNAMIC_CONTEXT_INJECTION]: The SKILL.md file was scanned for dynamic context injection markers (!command). No such patterns were found.
[PRIVILEGE_ESCALATION]: No commands requesting superuser privileges (sudo) or modifications to sensitive system directories were detected.
[PERSISTENCE]: The skill does not attempt to modify shell profiles, cron jobs, or startup services.

manimce-best-practices