Aegis Shield
Prompt-injection and data-exfiltration screening for untrusted text. Use before summarizing web/email/social content, before replying, and especially before writing anything to memory. Provides a safe memory append workflow (scan → lint → accept or quarantine).
Savant verdict: Pending analysis
Imported from the source hub; Savant's analysis is still running.
Live evaluation
Not evaluated yet. Workspaces can request a live evaluation.
Security assessment
Scan pending.
SKILL.md
This listing is enumerated but its package hasn't been fetched yet; it's queued by popularity.