The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data from the Bun documentation website (bun.com), creating a surface for indirect prompt injection.
Ingestion points: The agent is instructed to fetch content from https://bun.com/docs/llms.txt and various relative documentation paths (e.g., /docs/runtime/http/server.md) within SKILL.md.
Boundary markers: The prompt template provided for the sub-agent does not specify the use of delimiters or clear instructions to treat the fetched documentation strictly as data, which could allow instructions embedded in the docs to influence the agent.
Capability inventory: The skill utilizes the Task tool to spawn sub-agents for research, though it correctly sets the readonly parameter to true to limit the sub-agent's impact (SKILL.md).
Sanitization: There is no evidence of sanitization or filtering of the fetched documentation content before it is processed by the agent.

bun