The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests external, untrusted data in the form of SQL DDL and JSON schema/query files. This creates a surface for indirect prompt injection where an attacker could embed malicious instructions within the data to influence the agent's downstream actions or analysis.
Ingestion points: schema_analyzer.py (SQL and JSON inputs), index_optimizer.py (JSON inputs), and migration_generator.py (JSON inputs).
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potentially malicious content within these data files.
Capability inventory: The provided scripts allow for local file writing (via the --output argument), but do not perform network operations or execute arbitrary sub-processes.
Sanitization: schema_analyzer.py includes logic to strip SQL comments (-- and /* */), which mitigates some common indirect injection vectors, but does not sanitize the structural content of the DDL or JSON data.

database-designer