The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted user data from the conversation history to determine artifact types and routing logic, creating an indirect prompt injection surface.
Ingestion points: Step 1 in SKILL.md extracts signals, topics, concerns, and intent directly from the conversation history.
Boundary markers: No explicit delimiters or boundary markers are used to separate user instructions from data during analysis.
Capability inventory: The skill uses Read to access domain registry files, Write to save reports, and Skill to invoke the deepwiki skill.
Sanitization: There is no evidence of sanitization or filtering of the conversation signals before they are used to match domains or select routing strategies.
[EXTERNAL_DOWNLOADS]: The skill contains text instructions recommending the manual installation of a dependency (domain-registry) from the author's repository on GitHub.
Evidence: The SKILL.md file suggests using git clone https://github.com/mikeng-io/agent-skills if a dependency is missing. This refers to the vendor's own official infrastructure and is presented as a manual step for the user rather than an automated download.

context