The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data and has access to powerful tools.
Ingestion points: As described in SKILL.md, the skill is instructed to find and retrieve original declarations from manufacturer sites, program operator registries, and external databases like EC3.
Boundary markers: The instructions do not define specific delimiters or directives to the agent to ignore or isolate instructions that might be contained within the retrieved EPD documents.
Capability inventory: The skill's allowed-tools include Bash, Write, WebFetch, and WebSearch. It also manages the epd-library.csv file through a defined native validation and append process.
Sanitization: The skill does not specify any sanitization, validation, or filtering logic for the content it extracts from external EPDs before it is processed or recorded, creating a potential path for malicious data to influence agent behavior.

epd-research