The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a local Python script (scripts/check-output.py) used to validate that output deliverables contain specific markdown headers. This script uses standard libraries and operates only on local files provided as arguments, without requiring network access or administrative privileges.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests user requirements and local markdown files for validation. This data ingestion surface is not connected to any high-risk capabilities such as automated code execution or external data exfiltration, making the risk of indirect prompt injection negligible.\n- [METADATA_POISONING]: The skill's metadata and YAML frontmatter correctly describe its purpose and scope for CRM automation. No hidden or deceptive instructions were found in the metadata fields.

crm-integration