Release notes from agent-eval
Articles about Release notes from agent-eval
Merge pull request #978 from tangle-network/release/v0.213.0
chore(release): 0.213.0
V0.211.1: chore(release): 0.211.1 (#966)
chore(release): 0.211.1 Release-time preparation from merged main. Feature PRs do not carry version or changelog edits. A maintainer must approve the bot-created Actions workflow run before its checks can run.
V0.203.4: fix(deps): publish the 0.203 line on agent-core 0.10.2 (0.203.4) (#971)
fix(deps): publish the 0.203 line on agent-core 0.10.2 (0.203.4) v0.203.3 (#970) did not publish. Its packed-consumer check found two agent-interface copies, because agent-core 0.10.3 (2026-10-06) moved to agent-interface 3 in a patch release, and the 0.203 line's ^0.10.2 now resolves it beside the consumer's interface 2.
V0.203.3: feat(meta-eval): backport the judge gate to 0.203 (0.203.3) (#970)
feat(meta-eval): backport the judge gate to 0.203 (0.203.3) GTM pins agent-runtime 0.291.0 and agent-knowledge 18.0.0, whose agent-eval peer windows end below 0.204. The judge gate shipped in 0.210/0.211 (#961, #963), so GTM's adoption (#1445) broke its peer-floor ship check and was reverted (#1448). No agent-runtime or agent-knowledge release on GTM's interface-2 cohort admits agent-eval 0.211.
V0.211.0: chore(release): 0.211.0 (#964)
chore(release): 0.211.0 Release-time preparation from merged main using node scripts/prepare-release.mjs 0.211.0, matching the repository-owned Prepare Release workflow; uv lock --check passes. No feature or dependency-resolution changes. Carries the veto judge mode (#963). Publishing is the v0.211.0 tag on the merge commit.
V0.210.0: chore(release): 0.210.0 (#962)
chore(release): 0.210.0 Release-time preparation from merged main using node scripts/prepare-release.mjs 0.210.0, matching the repository-owned Prepare Release workflow; uv lock --check passes. No feature or dependency-resolution changes.
V0.208.0: chore(release): 0.208.0 (#943)
chore(release): 0.208.0 Release-time preparation from merged main. Feature PRs do not carry version or changelog edits. A maintainer must approve the bot-created Actions workflow run before its checks can run.
Release v0.207.4 · tangle-network/agent-eval
chore(release): 0.207.4 (#937)
V0.207.3: chore(release): 0.207.3 (#933)
Co-authored-by: github-actions[bot]
V0.207.1: chore(release): 0.207.1 (#929)
Co-authored-by: github-actions[bot]
Release v0.207.0 · tangle-network/agent-eval
v0.207.0
V0.203.0: chore(release): 0.203.0 (#904)
Co-authored-by: github-actions[bot]
Release v0.202.2 · tangle-network/agent-eval
v0.202.2
Release v0.202.1 · tangle-network/agent-eval
v0.202.1
V0.194.0: chore(release): prepare agent-eval 0.194.0 (#856)
Ships #855: a capped search whose cells overspent their holds can reach its claim again (the projector admits holds, not records), and verifySearchClaim makes a closed search's claim again from the ledger alone; runSearch throws on a mismatch and returns claimVerification. Also carries #853 (docs). Minor: verifySearchClaim and SearchRunResult.claimVerification are new API. Package, Python client and uv.lock at 0.194.0, CHANGELOG dated, analyst dependency-lock digest refreshed.
V0.193.2: chore(release): prepare agent-eval 0.193.2 (#851)
0.193.1's release commit (#850) bumped package.json, pyproject.toml and init.py but missed clients/python/uv.lock's own agent-eval-rpc package entry. The publish workflow's verifyVersionLock caught it before the npm publish step ran: "release version mismatch: npm=0.193.1, pythonPackage=0.193.1, pythonFallback=0.193.1, uvRoot=0.193.0" (run 36179196402). 0.193.1 was never published to npm or PyPI. Same content as 0.193.1, with uv.lock's version correctly locked this time. Verified locally by re-running verifyVersionLock's exact logic
V0.193.0: chore(release): prepare agent-eval 0.193.0 (#844)
Package, Python client and lock at 0.193.0, CHANGELOG dated, analyst dependency-lock digest refreshed. Breaking minor: carries #843 (the search claim) and #842 (search summary).
V0.192.0: chore(release): prepare agent-eval 0.192.0 (#841)
Package, Python client and lock at 0.192.0, CHANGELOG dated, and the public analyst benchmark dependency-lock digest refreshed for the version bump. Breaking minor: carries #840 (the search kernel under runOptimization; selectParent is policy, selectionRankKey, compareRankKeys, ParentSelector and the GenerationCandidate lineage fields deleted).
V0.191.0: chore(release): prepare agent-eval 0.191.0 (#839)
Package, Python client and lock at 0.191.0, CHANGELOG dated, and the public analyst benchmark dependency-lock digest refreshed for the version bump. Breaking minor: carries #838 (search-ledger wire and shipper; the eval-run wire, ingestEvalRuns, diffRuns and the winner-only shippers deleted).
Release v0.190.1 · tangle-network/agent-eval
chore(release): prepare agent-eval 0.190.1 (#837)
Release v0.190.0 · tangle-network/agent-eval
agent-eval 0.190.0
Release v0.185.0 · tangle-network/agent-eval
v0.185.0: agent-failure diagnosis engine
Release v0.184.0 · tangle-network/agent-eval
Merge pull request #782 from tangle-network/codex/jev-benchmark-integ
Release v0.183.0 · tangle-network/agent-eval
agent-eval 0.183.0
Release v0.182.0 · tangle-network/agent-eval
agent-eval 0.182.0
Release v0.181.0 · tangle-network/agent-eval
chore(release): publish eval 0.181.0 (#751)