The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided notebooks and markdown files, which presents a surface for indirect prompt injection. A malicious file could contain instructions that influence the agent's behavior during the cleanup or sharing process.
Ingestion points: Files are read using nbformat.read and standard file operations in SKILL.md, common-scenarios.md, and notebook-streamlining.md.
Boundary markers: Absent; there are no specific markers or warnings to the agent to ignore instructions embedded in the project files.
Capability inventory: The skill utilizes shell commands (mkdir, cp, mv, rm, zip, tar) and Python file operations (open, shutil) to modify the filesystem.
Sanitization: Absent; while the skill filters specific cell types and metadata, it does not sanitize content against adversarial instructions.\n- [EXTERNAL_DOWNLOADS]: The skill guides the creation of dependency files (requirements.txt, environment.yml), which results in the download and installation of external packages from public repositories such as PyPI and Conda.\n- [COMMAND_EXECUTION]: The skill involves the execution of various shell commands for project organization, file searching, and packaging, including mkdir, cp, mv, rm, zip, and tar.

project-sharing