The Agent Skills Directory

[SAFE]: The skill performs legitimate data analysis and report generation using standard Python libraries such as pandas. All operations are consistent with its stated purpose of creating business reports from structured data files.
[PROMPT_INJECTION]: While the skill processes external data (Excel/CSV), there is no evidence of instructions attempting to bypass safety filters or override agent behavior.
[DATA_EXFILTRATION]: The skill accesses local files (data.xlsx, data.csv) for processing but does not include any commands to send data to external or untrusted domains.
[INDIRECT_PROMPT_INJECTION]: The skill has an inherent attack surface for indirect prompt injection by ingesting external data.
Ingestion points: Data is loaded from data.xlsx and data.csv using pd.read_excel and pd.read_csv in SKILL.md.
Boundary markers: No explicit delimiters or instructions to ignore embedded content are present in the snippets.
Capability inventory: The skill uses pandas for data processing and supports generating files in .pptx, .docx, .html, and .xlsx formats.
Sanitization: No data sanitization or validation logic is shown in the provided code blocks.

data-report-generator