The Agent Skills Directory

[SAFE]: The skill correctly implements a recursion guard by checking for the RESEARCH_SUBAGENT environment variable before spawning new panes, preventing potential infinite loops of agent creation.
[SAFE]: The instructions for spawning a background agent using herdr pane split are a legitimate use of environment orchestration and do not involve unauthorized command execution or privilege escalation.
[SAFE]: The skill is designed to ingest information from external primary sources (documentation, source code, APIs). While this constitutes an attack surface for indirect prompt injection, it is the primary intended function of the skill and targets high-trust sources.
Ingestion points: Official documentation, source code, and APIs (SKILL.md).
Boundary markers: Absent.
Capability inventory: Spawns background agents via herdr and writes Markdown files to the repository (SKILL.md).
Sanitization: Absent.

research