The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes protein sequences which represent a potential surface for indirect prompt injection if malicious instructions are embedded in the sequence strings and subsequently interpreted by an LLM.
Ingestion points: The sequence parameter in find_n_glycosylation_sequons (SKILL.md) and batch analysis functions in references/glycan_databases.md.
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat sequence data as non-executable text.
Capability inventory: The skill utilizes requests.get for network operations and provides shell commands for package installation.
Sanitization: Input sequences are converted to uppercase and validated against biological motifs (N-X-[S/T]), which naturally restricts the character set but does not provide formal instruction filtering.
[EXTERNAL_DOWNLOADS]: The skill references and fetches data from well-known scientific and academic repositories.
Fetches protein-glycan profiles from glyconnect.expasy.org (Swiss Institute of Bioinformatics).
Retrieves glycan metadata from the api.glytoucan.org repository.
Points users to prediction services at services.healthtech.dtu.dk.
[COMMAND_EXECUTION]: Provides a command to install the glycoshield Python package via pip to enable glycan shielding analysis functionality.

glycoengineering