The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from Amplitude APIs, including experiment descriptions and user feedback insights. 1. Ingestion points: User-supplied experiment identifiers and qualitative feedback data retrieved via Amplitude MCP tools (SKILL.md, Steps 0 and 7). 2. Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the processed data. 3. Capability inventory: The skill uses Amplitude MCP tools for searching, reading, and managing experiments; it does not possess capabilities for arbitrary shell execution or local file system modification. 4. Sanitization: There is no mention of sanitizing or filtering the content retrieved from external sources before it is analyzed by the agent. This surface is considered a low risk associated with the skill's primary purpose.
[SAFE]: The skill follows professional standards for experiment analysis, including Sample Ratio Mismatch (SRM) checks, statistical power analysis, and multiple validity flags to ensure data quality. All operations are scoped to the Amplitude vendor environment.

analyze-experiments