The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill utilizes the uv run command to execute local Python scripts, specifically scripts/sympy_compute.py and scripts/z3_solve.py. These scripts are intended to perform mathematical computations and logic proofs using the SymPy and Z3 libraries respectively. This is a standard functional pattern for computational math skills.
[INDIRECT_PROMPT_INJECTION]: The skill defines a decision tree for processing external mathematical problems. This represents an ingestion point for untrusted data (user-provided math problems). While there are no explicit sanitization routines or boundary markers mentioned in the markdown, the nature of the task (mathematical integration) carries a low risk for prompt injection attacks.
[METADATA_POISONING]: The instructional text includes snippets and references to textbooks sourced from "Z-Library." While this indicates the documentation is derived from an external shadow library, the text itself does not contain malicious instructions or hidden payloads targeting the agent's behavior.

integration-theory