The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill uses the load_from_pickle function for data persistence as documented in SKILL.md and references/importing_exporting.md. This function utilizes the Python pickle module, which is insecure and can allow for arbitrary code execution if used to load a maliciously crafted file.
[EXTERNAL_DOWNLOADS]: The load_from_usi function described in references/importing_exporting.md allows the skill to retrieve spectral data from external online repositories using Universal Spectrum Identifiers (USI).
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted spectral data from various file formats, creating a surface for indirect prompt injection attacks.
- Ingestion points: Functions for loading MGF, mzML, MSP, JSON, and USI formats in references/importing_exporting.md.
- Boundary markers: No delimiters or protective instructions are documented to distinguish data from instructions.
- Capability inventory: The skill has the capability to write to the file system (e.g., save_as_mgf) and perform network operations (load_from_usi).
- Sanitization: No sanitization or validation logic is presented for the data ingested from these external files.

matchms