Default Branch

7e9aae4f77 · Merge pull request #202 from vercel-labs/changeset-release/main · Updated 2026-09-18 13:03:48 +08:00

Branches

7a4b029980 · fix: nested eval directory with housekeeping and result checking · Updated 2026-09-03 22:30:29 +08:00    skills

5
1

95a5e159c7 · feat: add natural multi-turn interaction · Updated 2026-08-12 07:26:09 +08:00    skills

45
4

4c9bf92fe6 · fix: snapshot exact fixture starting states · Updated 2026-08-12 07:25:55 +08:00    skills

45
3

821cc726e9 · agent-eval: add judge() — agentic LLM judge on the plugin model · Updated 2026-06-27 08:21:44 +08:00    skills

72
1

fec27a21f2 · agent-eval judge: codebase() — judge explores the workspace · Updated 2026-06-26 23:37:57 +08:00    skills

75
3

c9bb1f3aaa · agent-eval judge: real e2e + fixes it surfaced · Updated 2026-06-26 22:38:12 +08:00    skills

75
2

07838e6c2a · Skip npm install without package json · Updated 2026-05-07 07:40:22 +08:00    skills

119
1

20f540aaa6 · [Kimi] Add Kimi CLI agent harness with AI Gateway + direct Moonshot support · Updated 2026-04-21 06:21:26 +08:00    skills

131
1

55e0ab49b9 · [OpenCode] Add binaryUrl and extraProviders options for unreleased models · Updated 2026-04-08 07:14:14 +08:00    skills

142
1

4ed561af01 · Skip missing validation scripts · Updated 2026-03-19 22:53:10 +08:00    skills

147
1

e42dbf7d4c · playground changeset · Updated 2026-02-10 06:07:03 +08:00    skills

199
0
Included

739f70ee0a · [Codex] Support reasoning effort via model string, fix Docker CA certs, and improve npm install reliability · Updated 2026-02-10 03:08:41 +08:00    skills

206
1

465fbac30b · fix: allow VERCEL_OIDC_TOKEN if AI_GATEWAY_API_KEY is not set · Updated 2026-02-09 02:48:18 +08:00    skills

209
0
Included

b5f8f87c65 · Fingerprint reuse was never persisting fingerprints to results · Updated 2026-02-08 14:03:44 +08:00    skills

212
1

555b5c44ed · Overhaul README with run-all, fingerprinting, classification, and housekeeping · Updated 2026-02-08 13:39:46 +08:00    skills

213
1

9572432d9b · Add changeset for run-all and related features · Updated 2026-02-08 13:27:12 +08:00    skills

215
1

d24c847930 · Make run-all the default when invoked with no arguments · Updated 2026-02-08 13:22:53 +08:00    skills

216
1

6eb8bc6a55 · Add run-all command with housekeeping · Updated 2026-02-08 13:17:24 +08:00    skills

217
1

a61ddd8f0c · Add failure classification and auto-retry support · Updated 2026-02-08 13:15:43 +08:00    skills

218
1

cbd44c12a4 · Add content fingerprinting and result reuse (#39) · Updated 2026-02-08 12:52:28 +08:00    skills

220
2