The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill implements a local caching mechanism in references/data-access.md using the pickle module. The use of pickle.load() for reading cached data is an unsafe deserialization pattern that could allow arbitrary code execution if the local cache file is replaced or modified by a malicious actor.
[INDIRECT_PROMPT_INJECTION]: The skill processes extensive bioinformatics data from the DrugBank database, creating an attack surface for indirect prompt injection where malicious instructions could be embedded in database fields.
Ingestion points: XML database parsing as described in references/data-access.md and keyword-based drug queries in references/drug-queries.md.
Boundary markers: The skill does not implement delimiters or explicit instructions to the agent to ignore potentially malicious content embedded within the pharmaceutical data.
Capability inventory: The skill includes complex analysis tools for chemical similarity, molecular fingerprints (references/chemical-analysis.md), and interaction network analysis (references/interactions.md), but lacks dangerous capabilities like arbitrary system execution or file writing.
Sanitization: There is no evidence of sanitization, filtering, or validation of the text content retrieved from the XML database before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill facilitates the download of database updates and API queries from the official DrugBank domain (go.drugbank.com). These network operations are standard for the skill's intended purpose and target a well-known service domain.

drugbank-database