The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses Node.js (tools/apexctl.mjs) and Python scripts (runtime/internal/python/*) to execute shell commands, specifically SQLcl (sql) and the APEX CLI (apex). These commands are essential to the skill's primary function of validating and importing Oracle APEX artifacts. The execution is bounded to the local environment and uses static tool paths.
[INDIRECT_PROMPT_INJECTION]: The skill's core purpose is to ingest functional requirements (untrusted data) and generate executable code. This creates a potential surface for indirect prompt injection. However, the skill implements a comprehensive 'Draft -> Critique -> Revision' orchestration loop and enforces strict security policies (e.g., authenticated access by default, CSRF protection) to mitigate the risks associated with processing external requirements documents.
[EXTERNAL_DOWNLOADS]: The skill references and fetches content from official Oracle GitHub repositories (e.g., github.com/oracle/skills). These references are used for documentation and release notes. As Oracle is the skill author and a known service provider, these downloads are considered safe for the intended development context.
[SAFE]: The skill maintains a high security posture through its 'AI Tooling Guard' and 'Global Preferences' policies, which enforce session protection, access control, and data escaping in all generated APEX applications.

apexlang