Skip to content

Sync the skill body: catalog size, and the linear-pricing guidance - #10

Merged
kev1n merged 1 commit into
mainfrom
couple-catalog-size
Aug 29, 2026
Merged

kev1n merged 1 commit into
mainfrom
couple-catalog-size

Conversation

@kev1n

@kev1n kev1n commented Aug 29, 2026

Copy link
Copy Markdown
Member

npm run sync-skills in the anyapi repository, which the gate here requires to match src/lib/agentSkill.ts exactly. Two changes, both already merged upstream.

Catalog size. "You reach hundreds of APIs" becomes "You reach 359 APIs". Upstream, the published catalog size is now bound to the SKU manifests in the same commit, so every AnyAPI surface reads one constant and a stale figure fails CI before merge instead of after deploy. This skill was one of ten surfaces still stating the size in prose - vague rather than stale, but the same defect: a second owner for a number the repository already computes. Live /SKILL.md has served this since the frontend deployed (getanyapi-com/anyapi#840).

Linear-pricing guidance. Already out of sync here before this change, for an unrelated reason. Upstream split flat and linear offers because quoting maxPer1kUsd for a linear offer overstates a real call by roughly two orders of magnitude - linkedin.search_posts_full reads as 137 where 1,000 results actually cost $1.37. That guidance had never been synced, so this PR carries it.

bash scripts/check.sh passes: skills checks passed (1 manifest-listed skill).

Two changes, both already merged in the anyapi repository and carried here by
`npm run sync-skills`, which the external gate requires to match exactly.

The catalog size is now stated rather than gestured at. "Hundreds of APIs" was
a second owner for a number the main repository already computes from its SKU
manifests, so the skill said something vaguer than it needed to and could not
go stale in a way anyone would notice. It reads the same constant every other
AnyAPI surface reads, and says 359.

The pricing section was already out of sync here for an unrelated reason: main
had split flat and linear offers, because quoting `maxPer1kUsd` for a linear
offer overstates a call by roughly two orders of magnitude. That guidance had
never been synced.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
@kev1n
kev1n merged commit f2a4471 into main Aug 29, 2026
1 check passed
@kev1n
kev1n deleted the couple-catalog-size branch August 29, 2026 00:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant