The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (docs, source code, APIs) and writes output to the repository, which is a surface for indirect prompt injection attacks.
Ingestion points: SKILL.md instructs the agent to investigate official docs, source code, specs, and first-party APIs.
Boundary markers: Absent. The instructions do not define delimiters or provide warnings to the agent to ignore instructions found in the researched materials.
Capability inventory: The agent is authorized to write findings to Markdown files within the repository (SKILL.md).
Sanitization: Absent. There are no instructions to sanitize or validate the content retrieved from external sources.
[NO_CODE]: The skill contains configuration and instructions but does not include any executable scripts or binary files.

research