The Agent Skills Directory

[COMMAND_EXECUTION]: Local Tool Integration
The skill utilizes Node.js (child_process.spawnSync) and Python (subprocess.run) to execute internal helper scripts located within its package, such as tools/apexctl.mjs and runtime/internal/python/validate_apexlang.py.
These tools are used for workspace probing, preflight checks, and validating APEXlang DSL artifacts against compiler metadata.
[PROMPT_INJECTION]: Indirect Prompt Injection Surface
The skill ingests untrusted data from local environment files, including database metadata, data models, and business specifications, to drive code generation.
It employs strict instruction hierarchies and boundary markers (e.g., triple backticks for SQL/PLSQL blocks) to mitigate the risk of the agent obeying malicious instructions embedded in source data.
[DATA_EXFILTRATION]: Local Network Operations
The skill performs HTTP/HTTPS requests to verify that the generated application pages render correctly in the runtime environment.
These requests default to localhost and are used for diagnostic purposes. The verifier includes a bypass for SSL certificate validation (rejectUnauthorized: false) to facilitate connections to local development servers.

apex