The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow that ingests external data from physical well log files (LAS) and seismic volumes (SEGY).
Ingestion points: The workflow utilizes lasio.read() and segyio.open() to load external datasets into the analysis pipeline.
Boundary markers: The skill relies on standard Python data structures; no specific LLM-targeted delimiters are used.
Capability inventory: The identified capabilities are limited to local file reading, mathematical transformations (Gassmann fluid substitution, AVO modeling), and data visualization. There are no network write or shell execution capabilities.
Sanitization: The skill demonstrates best practices by including physical range assertions (e.g., checking that P-wave velocity is between 1500 and 7000 m/s) to ensure data integrity during processing.
[EXTERNAL_DOWNLOADS]: The skill references several third-party libraries including lasio, welly, bruges, and segyio. These are well-established, open-source tools within the geophysics community and are used for their intended domain-specific purposes.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns or unverified script downloads were detected. All code snippets provided are static Python examples for local execution.

rock-physics-avo