The Agent Skills Directory

[SAFE]: The skill implements standard machine learning infrastructure patterns for orchestration, tracking, and validation. No malicious code, obfuscation, or unauthorized network operations were detected.
[INDIRECT_PROMPT_INJECTION]: The skill defines ingestion points for external datasets (CSV, Parquet) in scripts like references/feature-engineering.md and references/pipeline-orchestration.md. While it recommends and provides code for data validation checkpoints using Great Expectations to mitigate risks, the agent handles untrusted data which is an inherent surface for indirect injection.
[DYNAMIC_EXECUTION]: The provided templates in SKILL.md and references/training-pipelines.md use pickle and joblib for model serialization and artifact management. These are standard tools in the ML domain for the primary purpose of the skill, although they involve deserialization of data.

ml-pipeline