The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied facts and data to generate report artifacts, creating a potential surface for indirect prompt injection if the inputs contain malicious instructions.
Ingestion points: The workflow relies on "supplied facts" and "supplied inputs" (SKILL.md) to populate the report packages and outlines.
Boundary markers: While the skill mentions using Markdown and JSON structures for output, it lacks explicit instructions or delimiters to isolate and ignore potentially malicious instructions embedded within the source data.
Capability inventory: The skill can invoke shell-based tools including omh CLI utilities and Python scripts (pptx_read.py, pptx_render.py) to process data and generate files (SKILL.md).
Sanitization: No explicit logic for sanitizing, validating, or escaping content from the "supplied facts" is described in the skill instructions.
[COMMAND_EXECUTION]: The skill workflow involves the execution of various command-line utilities to manage report state and artifacts.
The instructions specify the use of omh runtime record, omh achievements export, pptx_read.py, and pptx_render.py. These tools are documented as integrated platform utilities provided by the "oh-my-hermes" maintainers for the Hermes execution environment.

omh-report-package