The Agent Skills Directory

[SAFE]: The research workflow is logically sound and includes quality control measures like primary source verification and adversarial review.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface due to its web-research function. (1) Ingestion points: Web search and primary sources. (2) Boundary markers: Narrow agent briefs are used to scope sub-agent tasks. (3) Capability inventory: Web search, sub-agent spawning, and markdown file writing. (4) Sanitization: No explicit sanitization of external data is mentioned. This surface is inherent to the skill's research purpose and is considered safe.
[NO_CODE]: The skill contains only markdown instructions and configuration metadata; it does not ship with any scripts, binaries, or executable code.

deep-research