Sign in
Skills

Iteratively improves a SKILL.md via LLM passes guided by the review pipeline; returns the diff and revised content.

post/experimental/skills/improve

Headers

Authorizationstring

Request body

contentstring required
modelstring

Model override to use for the improve passes instead of the default model

max-iterationsinteger

Maximum number of improve iterations (1-10, default 3)

Response

Default Response

originalContentstring required
improvedContentstring required
diffstring required
reviewSummarystring required
summaryOfChangesstring required
originalScoreinteger nullable required
finalScoreinteger nullable required
iterationsinteger required
creditsnumber required

Credits charged for this improvement (0 if nothing was billed).