The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill is vulnerable to indirect prompt injection because it ingests untrusted data from the local filesystem and processes it using high-capability tools and subagents.
Ingestion points: The skill utilizes Read, Glob, and Grep tools to pull external file content into the agent context.
Boundary markers: The skill definition lacks delimiters or explicit instructions to ignore embedded commands within the read files.
Capability inventory: The skill has access to Bash, Task, and the ability to spawn Explore subagents, providing a significant surface for exploitation if an injection occurs.
Sanitization: There is no evidence of sanitization, escaping, or validation of the content retrieved from the data sources before it is passed to the LLM or subagents.

breadth-reader