docx (anthropics/skills) vs verification-before-completion (obra/superpowers)
docx (anthropics/skills)
21.4 / 30
SituationalSafe
anthropics docx: same results as no skill on three Word tasks, plus working files left behind
- SOTAPass
verification-before-completion (obra/superpowers)
25.0 / 30
SituationalSafe
verification-before-completion: all 18 runs verify before claiming done, with or without the skill (5.0 / 5.0 / 5.0); the skill adds ~20% tokens
- SOTAPass