The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides a script scripts/dispersion_analysis.py that ingests external data from user-specified files. This creates a potential surface for indirect prompt injection if an attacker provides a malicious data file, although the risk is mitigated by the script's use of numpy.loadtxt which strictly parses numeric data.\n
Ingestion points: The load_model_file function in scripts/dispersion_analysis.py reads data from a path provided via the --input argument.\n
Boundary markers: No explicit delimiters or instructions to ignore embedded content are used during file ingestion.\n
Capability inventory: The skill possesses file-writing capabilities through the save_dispersion function, which uses numpy.savetxt to write results to a user-defined path.\n
Sanitization: The ingestion process uses numpy.loadtxt, which effectively sanitizes input by requiring it to conform to a specific numeric array structure, failing on arbitrary text or code.\n- [DYNAMIC_EXECUTION]: The skill utilizes numba for Just-In-Time (JIT) compilation to accelerate seismic dispersion calculations. While this is a standard and expected practice in high-performance scientific Python applications, it involves the generation and execution of code at runtime.

disba