Create detailed phase plan (PLAN.md) with verification loop
The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
The skill workflow ingests arbitrary web pages via WebFetch and optional paths/globs, but no monitored outsider submission feed or public search source is unconditionally consumed.
open-gsd/gsd-core
99bbbac
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.