The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests untrusted data from an Oracle database (source code, identifiers, and comments) through commands such as pythia src, pythia grep, and pythia sql. This data is processed by the agent, creating a surface for indirect prompt injection where adversarial instructions embedded in the database could influence agent behavior.
Ingestion points: Database content retrieved via pythia src, pythia grep, and pythia sql in SKILL.md.
Boundary markers: No specific delimiters or instructions to ignore embedded content were identified.
Capability inventory: The skill performs database queries and reads schema definitions via the pythia CLI tool.
Sanitization: No evidence of sanitization or filtering of database-sourced content is present.
[COMMAND_EXECUTION]: The skill utilizes a CLI utility named pythia to interact with the database. Specifically, the pythia sql command enables the execution of free-form SQL queries. Although the instructions restrict usage to SELECT and WITH statements, it allows for arbitrary read-access queries against any data the connected user is authorized to view.

plsql-explore