@cursor/july 0.1.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/AGENTS.md +290 -0
- package/LICENSE.md +3 -0
- package/README.md +1233 -0
- package/dist/ab.d.ts +290 -0
- package/dist/ab.d.ts.map +1 -0
- package/dist/ab.js +386 -0
- package/dist/bin/agent-serve.d.ts +15 -0
- package/dist/bin/agent-serve.d.ts.map +1 -0
- package/dist/bin/agent-serve.js +754 -0
- package/dist/channels/github/api.d.ts +101 -0
- package/dist/channels/github/api.d.ts.map +1 -0
- package/dist/channels/github/api.js +229 -0
- package/dist/channels/github/auth.d.ts +67 -0
- package/dist/channels/github/auth.d.ts.map +1 -0
- package/dist/channels/github/auth.js +239 -0
- package/dist/channels/github/binding.d.ts +73 -0
- package/dist/channels/github/binding.d.ts.map +1 -0
- package/dist/channels/github/binding.js +114 -0
- package/dist/channels/github/defaults.d.ts +24 -0
- package/dist/channels/github/defaults.d.ts.map +1 -0
- package/dist/channels/github/defaults.js +100 -0
- package/dist/channels/github/events.d.ts +18 -0
- package/dist/channels/github/events.d.ts.map +1 -0
- package/dist/channels/github/events.js +59 -0
- package/dist/channels/github/forward.d.ts +137 -0
- package/dist/channels/github/forward.d.ts.map +1 -0
- package/dist/channels/github/forward.js +271 -0
- package/dist/channels/github/github-channel.d.ts +127 -0
- package/dist/channels/github/github-channel.d.ts.map +1 -0
- package/dist/channels/github/github-channel.js +1688 -0
- package/dist/channels/github/index.d.ts +52 -0
- package/dist/channels/github/index.d.ts.map +1 -0
- package/dist/channels/github/index.js +51 -0
- package/dist/channels/github/replay.d.ts +115 -0
- package/dist/channels/github/replay.d.ts.map +1 -0
- package/dist/channels/github/replay.js +282 -0
- package/dist/channels/github/state.d.ts +26 -0
- package/dist/channels/github/state.d.ts.map +1 -0
- package/dist/channels/github/state.js +50 -0
- package/dist/channels/github/types.d.ts +385 -0
- package/dist/channels/github/types.d.ts.map +1 -0
- package/dist/channels/github/types.js +1 -0
- package/dist/channels/slack/api.d.ts +117 -0
- package/dist/channels/slack/api.d.ts.map +1 -0
- package/dist/channels/slack/api.js +437 -0
- package/dist/channels/slack/approvals.d.ts +32 -0
- package/dist/channels/slack/approvals.d.ts.map +1 -0
- package/dist/channels/slack/approvals.js +157 -0
- package/dist/channels/slack/assistant.d.ts +23 -0
- package/dist/channels/slack/assistant.d.ts.map +1 -0
- package/dist/channels/slack/assistant.js +67 -0
- package/dist/channels/slack/auth.d.ts +40 -0
- package/dist/channels/slack/auth.d.ts.map +1 -0
- package/dist/channels/slack/auth.js +71 -0
- package/dist/channels/slack/channel-watch.d.ts +46 -0
- package/dist/channels/slack/channel-watch.d.ts.map +1 -0
- package/dist/channels/slack/channel-watch.js +180 -0
- package/dist/channels/slack/constants.d.ts +17 -0
- package/dist/channels/slack/constants.d.ts.map +1 -0
- package/dist/channels/slack/constants.js +27 -0
- package/dist/channels/slack/credentials.d.ts +35 -0
- package/dist/channels/slack/credentials.d.ts.map +1 -0
- package/dist/channels/slack/credentials.js +88 -0
- package/dist/channels/slack/cursor-account.d.ts +76 -0
- package/dist/channels/slack/cursor-account.d.ts.map +1 -0
- package/dist/channels/slack/cursor-account.js +88 -0
- package/dist/channels/slack/defaults.d.ts +32 -0
- package/dist/channels/slack/defaults.d.ts.map +1 -0
- package/dist/channels/slack/defaults.js +380 -0
- package/dist/channels/slack/dispatch.d.ts +43 -0
- package/dist/channels/slack/dispatch.d.ts.map +1 -0
- package/dist/channels/slack/dispatch.js +185 -0
- package/dist/channels/slack/doctor.d.ts +24 -0
- package/dist/channels/slack/doctor.d.ts.map +1 -0
- package/dist/channels/slack/doctor.js +115 -0
- package/dist/channels/slack/event-id-dedupe.d.ts +10 -0
- package/dist/channels/slack/event-id-dedupe.d.ts.map +1 -0
- package/dist/channels/slack/event-id-dedupe.js +23 -0
- package/dist/channels/slack/inbound.d.ts +41 -0
- package/dist/channels/slack/inbound.d.ts.map +1 -0
- package/dist/channels/slack/inbound.js +166 -0
- package/dist/channels/slack/index.d.ts +45 -0
- package/dist/channels/slack/index.d.ts.map +1 -0
- package/dist/channels/slack/index.js +44 -0
- package/dist/channels/slack/init.d.ts +32 -0
- package/dist/channels/slack/init.d.ts.map +1 -0
- package/dist/channels/slack/init.js +82 -0
- package/dist/channels/slack/interactive.d.ts +19 -0
- package/dist/channels/slack/interactive.d.ts.map +1 -0
- package/dist/channels/slack/interactive.js +142 -0
- package/dist/channels/slack/links.d.ts +10 -0
- package/dist/channels/slack/links.d.ts.map +1 -0
- package/dist/channels/slack/links.js +19 -0
- package/dist/channels/slack/live-delivery.d.ts +75 -0
- package/dist/channels/slack/live-delivery.d.ts.map +1 -0
- package/dist/channels/slack/live-delivery.js +143 -0
- package/dist/channels/slack/log.d.ts +25 -0
- package/dist/channels/slack/log.d.ts.map +1 -0
- package/dist/channels/slack/log.js +47 -0
- package/dist/channels/slack/manifest.d.ts +57 -0
- package/dist/channels/slack/manifest.d.ts.map +1 -0
- package/dist/channels/slack/manifest.js +110 -0
- package/dist/channels/slack/pr-ref.d.ts +18 -0
- package/dist/channels/slack/pr-ref.d.ts.map +1 -0
- package/dist/channels/slack/pr-ref.js +36 -0
- package/dist/channels/slack/setup.d.ts +14 -0
- package/dist/channels/slack/setup.d.ts.map +1 -0
- package/dist/channels/slack/setup.js +86 -0
- package/dist/channels/slack/slack-channel.d.ts +15 -0
- package/dist/channels/slack/slack-channel.d.ts.map +1 -0
- package/dist/channels/slack/slack-channel.js +281 -0
- package/dist/channels/slack/socket-mode.d.ts +26 -0
- package/dist/channels/slack/socket-mode.d.ts.map +1 -0
- package/dist/channels/slack/socket-mode.js +116 -0
- package/dist/channels/slack/stream.d.ts +26 -0
- package/dist/channels/slack/stream.d.ts.map +1 -0
- package/dist/channels/slack/stream.js +39 -0
- package/dist/channels/slack/tool-approval-events.d.ts +33 -0
- package/dist/channels/slack/tool-approval-events.d.ts.map +1 -0
- package/dist/channels/slack/tool-approval-events.js +129 -0
- package/dist/channels/slack/types.d.ts +326 -0
- package/dist/channels/slack/types.d.ts.map +1 -0
- package/dist/channels/slack/types.js +4 -0
- package/dist/channels.d.ts +116 -0
- package/dist/channels.d.ts.map +1 -0
- package/dist/channels.js +294 -0
- package/dist/connections.d.ts +75 -0
- package/dist/connections.d.ts.map +1 -0
- package/dist/connections.js +75 -0
- package/dist/docs/404.html +22 -0
- package/dist/docs/ab.html +87 -0
- package/dist/docs/assets/ab.md.COdXkces.js +63 -0
- package/dist/docs/assets/ab.md.COdXkces.lean.js +1 -0
- package/dist/docs/assets/app.DqfFEmJd.js +1 -0
- package/dist/docs/assets/building-with-agents.md.CJCtZCyi.js +13 -0
- package/dist/docs/assets/building-with-agents.md.CJCtZCyi.lean.js +1 -0
- package/dist/docs/assets/chunks/@localSearchIndexroot.CcVk1uKq.js +1 -0
- package/dist/docs/assets/chunks/VPLocalSearchBox.BaLEdS15.js +9 -0
- package/dist/docs/assets/chunks/framework.CAZyNGu9.js +19 -0
- package/dist/docs/assets/chunks/theme.CZRvu_0q.js +2 -0
- package/dist/docs/assets/concepts.md.Cfb9b-k1.js +4 -0
- package/dist/docs/assets/concepts.md.Cfb9b-k1.lean.js +1 -0
- package/dist/docs/assets/deployment.md.Dx1TYNk5.js +14 -0
- package/dist/docs/assets/deployment.md.Dx1TYNk5.lean.js +1 -0
- package/dist/docs/assets/evals.md.DPZ_MAnI.js +85 -0
- package/dist/docs/assets/evals.md.DPZ_MAnI.lean.js +1 -0
- package/dist/docs/assets/guides_agent-to-agent.md.CrtrsySy.js +10 -0
- package/dist/docs/assets/guides_agent-to-agent.md.CrtrsySy.lean.js +1 -0
- package/dist/docs/assets/guides_cloud-runtime.md.CYlNMTNp.js +9 -0
- package/dist/docs/assets/guides_cloud-runtime.md.CYlNMTNp.lean.js +1 -0
- package/dist/docs/assets/guides_github.md.DwbKhCeS.js +20 -0
- package/dist/docs/assets/guides_github.md.DwbKhCeS.lean.js +1 -0
- package/dist/docs/assets/guides_human-in-the-loop.md.Dvuctx7s.js +28 -0
- package/dist/docs/assets/guides_human-in-the-loop.md.Dvuctx7s.lean.js +1 -0
- package/dist/docs/assets/guides_slack.md.bv41fHfW.js +37 -0
- package/dist/docs/assets/guides_slack.md.bv41fHfW.lean.js +1 -0
- package/dist/docs/assets/guides_webhooks.md.hFTik3lf.js +50 -0
- package/dist/docs/assets/guides_webhooks.md.hFTik3lf.lean.js +1 -0
- package/dist/docs/assets/hillclimbing.md.DHNast08.js +4 -0
- package/dist/docs/assets/hillclimbing.md.DHNast08.lean.js +1 -0
- package/dist/docs/assets/index.md.BPKcj5AI.js +20 -0
- package/dist/docs/assets/index.md.BPKcj5AI.lean.js +1 -0
- package/dist/docs/assets/inter-italic-cyrillic-ext.r48I6akx.woff2 +0 -0
- package/dist/docs/assets/inter-italic-cyrillic.By2_1cv3.woff2 +0 -0
- package/dist/docs/assets/inter-italic-greek-ext.1u6EdAuj.woff2 +0 -0
- package/dist/docs/assets/inter-italic-greek.DJ8dCoTZ.woff2 +0 -0
- package/dist/docs/assets/inter-italic-latin-ext.CN1xVJS-.woff2 +0 -0
- package/dist/docs/assets/inter-italic-latin.C2AdPX0b.woff2 +0 -0
- package/dist/docs/assets/inter-italic-vietnamese.BSbpV94h.woff2 +0 -0
- package/dist/docs/assets/inter-roman-cyrillic-ext.BBPuwvHQ.woff2 +0 -0
- package/dist/docs/assets/inter-roman-cyrillic.C5lxZ8CY.woff2 +0 -0
- package/dist/docs/assets/inter-roman-greek-ext.CqjqNYQ-.woff2 +0 -0
- package/dist/docs/assets/inter-roman-greek.BBVDIX6e.woff2 +0 -0
- package/dist/docs/assets/inter-roman-latin-ext.4ZJIpNVo.woff2 +0 -0
- package/dist/docs/assets/inter-roman-latin.Di8DUHzh.woff2 +0 -0
- package/dist/docs/assets/inter-roman-vietnamese.BjW4sHH5.woff2 +0 -0
- package/dist/docs/assets/quickstart.md.tVPiGK_L.js +35 -0
- package/dist/docs/assets/quickstart.md.tVPiGK_L.lean.js +1 -0
- package/dist/docs/assets/reference_agent-config.md.Bpd7HQwf.js +34 -0
- package/dist/docs/assets/reference_agent-config.md.Bpd7HQwf.lean.js +1 -0
- package/dist/docs/assets/reference_channels.md.D7JTR03W.js +53 -0
- package/dist/docs/assets/reference_channels.md.D7JTR03W.lean.js +1 -0
- package/dist/docs/assets/reference_cli.md.DA730zCu.js +33 -0
- package/dist/docs/assets/reference_cli.md.DA730zCu.lean.js +1 -0
- package/dist/docs/assets/reference_connections.md.C3vNH_DE.js +26 -0
- package/dist/docs/assets/reference_connections.md.C3vNH_DE.lean.js +1 -0
- package/dist/docs/assets/reference_hooks.md.BCEc3MyM.js +12 -0
- package/dist/docs/assets/reference_hooks.md.BCEc3MyM.lean.js +1 -0
- package/dist/docs/assets/reference_http-api.md.DBAahtdz.js +11 -0
- package/dist/docs/assets/reference_http-api.md.DBAahtdz.lean.js +1 -0
- package/dist/docs/assets/reference_instructions.md.BC05LEQ8.js +12 -0
- package/dist/docs/assets/reference_instructions.md.BC05LEQ8.lean.js +1 -0
- package/dist/docs/assets/reference_playground.md.CPZhfYaO.js +1 -0
- package/dist/docs/assets/reference_playground.md.CPZhfYaO.lean.js +1 -0
- package/dist/docs/assets/reference_project-layout.md.Bd_CKtNS.js +31 -0
- package/dist/docs/assets/reference_project-layout.md.Bd_CKtNS.lean.js +1 -0
- package/dist/docs/assets/reference_schedules.md.D7qijxLk.js +82 -0
- package/dist/docs/assets/reference_schedules.md.D7qijxLk.lean.js +1 -0
- package/dist/docs/assets/reference_sessions.md.DdjWabnN.js +8 -0
- package/dist/docs/assets/reference_sessions.md.DdjWabnN.lean.js +1 -0
- package/dist/docs/assets/reference_skills.md.VQnlBT3Q.js +15 -0
- package/dist/docs/assets/reference_skills.md.VQnlBT3Q.lean.js +1 -0
- package/dist/docs/assets/reference_subagents.md.CIRAVcPK.js +10 -0
- package/dist/docs/assets/reference_subagents.md.CIRAVcPK.lean.js +1 -0
- package/dist/docs/assets/reference_tools.md.DF5kwlt0.js +46 -0
- package/dist/docs/assets/reference_tools.md.DF5kwlt0.lean.js +1 -0
- package/dist/docs/assets/scaffolding-agents.md.CyYfWGdc.js +1 -0
- package/dist/docs/assets/scaffolding-agents.md.CyYfWGdc.lean.js +1 -0
- package/dist/docs/assets/style.Bj-IXa5W.css +1 -0
- package/dist/docs/assets/troubleshooting.md.DkovjS7h.js +1 -0
- package/dist/docs/assets/troubleshooting.md.DkovjS7h.lean.js +1 -0
- package/dist/docs/building-with-agents.html +37 -0
- package/dist/docs/concepts.html +28 -0
- package/dist/docs/deployment.html +38 -0
- package/dist/docs/evals.html +109 -0
- package/dist/docs/guides/agent-to-agent.html +34 -0
- package/dist/docs/guides/cloud-runtime.html +33 -0
- package/dist/docs/guides/github.html +44 -0
- package/dist/docs/guides/human-in-the-loop.html +52 -0
- package/dist/docs/guides/slack.html +61 -0
- package/dist/docs/guides/webhooks.html +74 -0
- package/dist/docs/hashmap.json +1 -0
- package/dist/docs/hillclimbing.html +28 -0
- package/dist/docs/index.html +44 -0
- package/dist/docs/quickstart.html +59 -0
- package/dist/docs/reference/agent-config.html +58 -0
- package/dist/docs/reference/channels.html +77 -0
- package/dist/docs/reference/cli.html +57 -0
- package/dist/docs/reference/connections.html +50 -0
- package/dist/docs/reference/hooks.html +36 -0
- package/dist/docs/reference/http-api.html +35 -0
- package/dist/docs/reference/instructions.html +36 -0
- package/dist/docs/reference/playground.html +25 -0
- package/dist/docs/reference/project-layout.html +55 -0
- package/dist/docs/reference/schedules.html +106 -0
- package/dist/docs/reference/sessions.html +32 -0
- package/dist/docs/reference/skills.html +39 -0
- package/dist/docs/reference/subagents.html +34 -0
- package/dist/docs/reference/tools.html +70 -0
- package/dist/docs/scaffolding-agents.html +25 -0
- package/dist/docs/troubleshooting.html +25 -0
- package/dist/docs/vp-icons.css +0 -0
- package/dist/evals.d.ts +312 -0
- package/dist/evals.d.ts.map +1 -0
- package/dist/evals.js +254 -0
- package/dist/hooks.d.ts +23 -0
- package/dist/hooks.d.ts.map +1 -0
- package/dist/hooks.js +23 -0
- package/dist/index.d.ts +71 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +101 -0
- package/dist/internal/ab-collector.d.ts +42 -0
- package/dist/internal/ab-collector.d.ts.map +1 -0
- package/dist/internal/ab-collector.js +153 -0
- package/dist/internal/ab-fold.d.ts +36 -0
- package/dist/internal/ab-fold.d.ts.map +1 -0
- package/dist/internal/ab-fold.js +171 -0
- package/dist/internal/ab-snapshot.d.ts +70 -0
- package/dist/internal/ab-snapshot.d.ts.map +1 -0
- package/dist/internal/ab-snapshot.js +208 -0
- package/dist/internal/approval-gate.d.ts +34 -0
- package/dist/internal/approval-gate.d.ts.map +1 -0
- package/dist/internal/approval-gate.js +85 -0
- package/dist/internal/auth.d.ts +8 -0
- package/dist/internal/auth.d.ts.map +1 -0
- package/dist/internal/auth.js +27 -0
- package/dist/internal/authored-alias-hooks.d.ts +31 -0
- package/dist/internal/authored-alias-hooks.d.ts.map +1 -0
- package/dist/internal/authored-alias-hooks.js +43 -0
- package/dist/internal/authored-loaders.d.ts +25 -0
- package/dist/internal/authored-loaders.d.ts.map +1 -0
- package/dist/internal/authored-loaders.js +129 -0
- package/dist/internal/bounded-int.d.ts +11 -0
- package/dist/internal/bounded-int.d.ts.map +1 -0
- package/dist/internal/bounded-int.js +13 -0
- package/dist/internal/brand.d.ts +6 -0
- package/dist/internal/brand.d.ts.map +1 -0
- package/dist/internal/brand.js +17 -0
- package/dist/internal/chat-client.d.ts +19 -0
- package/dist/internal/chat-client.d.ts.map +1 -0
- package/dist/internal/chat-client.js +327 -0
- package/dist/internal/cli-ax.d.ts +58 -0
- package/dist/internal/cli-ax.d.ts.map +1 -0
- package/dist/internal/cli-ax.js +345 -0
- package/dist/internal/cli-cursor.d.ts +20 -0
- package/dist/internal/cli-cursor.d.ts.map +1 -0
- package/dist/internal/cli-cursor.js +227 -0
- package/dist/internal/cli-github.d.ts +47 -0
- package/dist/internal/cli-github.d.ts.map +1 -0
- package/dist/internal/cli-github.js +644 -0
- package/dist/internal/cli-slack.d.ts +18 -0
- package/dist/internal/cli-slack.d.ts.map +1 -0
- package/dist/internal/cli-slack.js +144 -0
- package/dist/internal/cloud-merge.d.ts +23 -0
- package/dist/internal/cloud-merge.d.ts.map +1 -0
- package/dist/internal/cloud-merge.js +28 -0
- package/dist/internal/cron.d.ts +17 -0
- package/dist/internal/cron.d.ts.map +1 -0
- package/dist/internal/cron.js +60 -0
- package/dist/internal/cursor/account-mcp.d.ts +95 -0
- package/dist/internal/cursor/account-mcp.d.ts.map +1 -0
- package/dist/internal/cursor/account-mcp.js +342 -0
- package/dist/internal/cursor/backend-client.d.ts +118 -0
- package/dist/internal/cursor/backend-client.d.ts.map +1 -0
- package/dist/internal/cursor/backend-client.js +477 -0
- package/dist/internal/cursor/credentials.d.ts +112 -0
- package/dist/internal/cursor/credentials.d.ts.map +1 -0
- package/dist/internal/cursor/credentials.js +329 -0
- package/dist/internal/cursor-event-relay.d.ts +111 -0
- package/dist/internal/cursor-event-relay.d.ts.map +1 -0
- package/dist/internal/cursor-event-relay.js +181 -0
- package/dist/internal/cursor-relay-core.d.ts +113 -0
- package/dist/internal/cursor-relay-core.d.ts.map +1 -0
- package/dist/internal/cursor-relay-core.js +364 -0
- package/dist/internal/cursor-slack-relay.d.ts +94 -0
- package/dist/internal/cursor-slack-relay.d.ts.map +1 -0
- package/dist/internal/cursor-slack-relay.js +171 -0
- package/dist/internal/describe-error.d.ts +3 -0
- package/dist/internal/describe-error.d.ts.map +1 -0
- package/dist/internal/describe-error.js +4 -0
- package/dist/internal/dev-mode.d.ts +13 -0
- package/dist/internal/dev-mode.d.ts.map +1 -0
- package/dist/internal/dev-mode.js +17 -0
- package/dist/internal/discovery.d.ts +55 -0
- package/dist/internal/discovery.d.ts.map +1 -0
- package/dist/internal/discovery.js +1427 -0
- package/dist/internal/distribution.d.ts +47 -0
- package/dist/internal/distribution.d.ts.map +1 -0
- package/dist/internal/distribution.js +82 -0
- package/dist/internal/docs-site.d.ts +28 -0
- package/dist/internal/docs-site.d.ts.map +1 -0
- package/dist/internal/docs-site.js +169 -0
- package/dist/internal/duration.d.ts +8 -0
- package/dist/internal/duration.d.ts.map +1 -0
- package/dist/internal/duration.js +35 -0
- package/dist/internal/eval-run-store.d.ts +81 -0
- package/dist/internal/eval-run-store.d.ts.map +1 -0
- package/dist/internal/eval-run-store.js +396 -0
- package/dist/internal/eval-runner.d.ts +76 -0
- package/dist/internal/eval-runner.d.ts.map +1 -0
- package/dist/internal/eval-runner.js +336 -0
- package/dist/internal/event-log.d.ts +51 -0
- package/dist/internal/event-log.d.ts.map +1 -0
- package/dist/internal/event-log.js +167 -0
- package/dist/internal/event-mapper.d.ts +46 -0
- package/dist/internal/event-mapper.d.ts.map +1 -0
- package/dist/internal/event-mapper.js +226 -0
- package/dist/internal/frontmatter.d.ts +12 -0
- package/dist/internal/frontmatter.d.ts.map +1 -0
- package/dist/internal/frontmatter.js +61 -0
- package/dist/internal/github-fanout.d.ts +32 -0
- package/dist/internal/github-fanout.d.ts.map +1 -0
- package/dist/internal/github-fanout.js +122 -0
- package/dist/internal/handleAgentServeTrigger.d.ts +59 -0
- package/dist/internal/handleAgentServeTrigger.d.ts.map +1 -0
- package/dist/internal/handleAgentServeTrigger.js +115 -0
- package/dist/internal/host-platforms.d.ts +40 -0
- package/dist/internal/host-platforms.d.ts.map +1 -0
- package/dist/internal/host-platforms.js +113 -0
- package/dist/internal/http-channel.d.ts +54 -0
- package/dist/internal/http-channel.d.ts.map +1 -0
- package/dist/internal/http-channel.js +382 -0
- package/dist/internal/ids.d.ts +19 -0
- package/dist/internal/ids.d.ts.map +1 -0
- package/dist/internal/ids.js +34 -0
- package/dist/internal/init-project.d.ts +22 -0
- package/dist/internal/init-project.d.ts.map +1 -0
- package/dist/internal/init-project.js +125 -0
- package/dist/internal/json-dir-store.d.ts +32 -0
- package/dist/internal/json-dir-store.d.ts.map +1 -0
- package/dist/internal/json-dir-store.js +100 -0
- package/dist/internal/json-schema.d.ts +18 -0
- package/dist/internal/json-schema.d.ts.map +1 -0
- package/dist/internal/json-schema.js +53 -0
- package/dist/internal/log-ring.d.ts +23 -0
- package/dist/internal/log-ring.d.ts.map +1 -0
- package/dist/internal/log-ring.js +42 -0
- package/dist/internal/map-pool.d.ts +6 -0
- package/dist/internal/map-pool.d.ts.map +1 -0
- package/dist/internal/map-pool.js +41 -0
- package/dist/internal/mcp-endpoint.d.ts +59 -0
- package/dist/internal/mcp-endpoint.d.ts.map +1 -0
- package/dist/internal/mcp-endpoint.js +355 -0
- package/dist/internal/mcp-host.d.ts +28 -0
- package/dist/internal/mcp-host.d.ts.map +1 -0
- package/dist/internal/mcp-host.js +162 -0
- package/dist/internal/net.d.ts +5 -0
- package/dist/internal/net.d.ts.map +1 -0
- package/dist/internal/net.js +30 -0
- package/dist/internal/peer-connections.d.ts +44 -0
- package/dist/internal/peer-connections.d.ts.map +1 -0
- package/dist/internal/peer-connections.js +67 -0
- package/dist/internal/playground/static.d.ts +35 -0
- package/dist/internal/playground/static.d.ts.map +1 -0
- package/dist/internal/playground/static.js +162 -0
- package/dist/internal/playground/toolchain.d.ts +57 -0
- package/dist/internal/playground/toolchain.d.ts.map +1 -0
- package/dist/internal/playground/toolchain.js +226 -0
- package/dist/internal/reminder-runner.d.ts +62 -0
- package/dist/internal/reminder-runner.d.ts.map +1 -0
- package/dist/internal/reminder-runner.js +582 -0
- package/dist/internal/reminder-schedule.d.ts +22 -0
- package/dist/internal/reminder-schedule.d.ts.map +1 -0
- package/dist/internal/reminder-schedule.js +82 -0
- package/dist/internal/reminder-store.d.ts +58 -0
- package/dist/internal/reminder-store.d.ts.map +1 -0
- package/dist/internal/reminder-store.js +106 -0
- package/dist/internal/resolved-connections.d.ts +41 -0
- package/dist/internal/resolved-connections.d.ts.map +1 -0
- package/dist/internal/resolved-connections.js +73 -0
- package/dist/internal/review-comments.d.ts +50 -0
- package/dist/internal/review-comments.d.ts.map +1 -0
- package/dist/internal/review-comments.js +194 -0
- package/dist/internal/router.d.ts +33 -0
- package/dist/internal/router.d.ts.map +1 -0
- package/dist/internal/router.js +93 -0
- package/dist/internal/run-client.d.ts +49 -0
- package/dist/internal/run-client.d.ts.map +1 -0
- package/dist/internal/run-client.js +208 -0
- package/dist/internal/schedule-runner.d.ts +31 -0
- package/dist/internal/schedule-runner.d.ts.map +1 -0
- package/dist/internal/schedule-runner.js +126 -0
- package/dist/internal/sdk-runner.d.ts +41 -0
- package/dist/internal/sdk-runner.d.ts.map +1 -0
- package/dist/internal/sdk-runner.js +221 -0
- package/dist/internal/server.d.ts +25 -0
- package/dist/internal/server.d.ts.map +1 -0
- package/dist/internal/server.js +1374 -0
- package/dist/internal/session-engine.d.ts +287 -0
- package/dist/internal/session-engine.d.ts.map +1 -0
- package/dist/internal/session-engine.js +1377 -0
- package/dist/internal/session-store.d.ts +29 -0
- package/dist/internal/session-store.d.ts.map +1 -0
- package/dist/internal/session-store.js +140 -0
- package/dist/internal/session-urls.d.ts +10 -0
- package/dist/internal/session-urls.d.ts.map +1 -0
- package/dist/internal/session-urls.js +25 -0
- package/dist/internal/stream-progress.d.ts +14 -0
- package/dist/internal/stream-progress.d.ts.map +1 -0
- package/dist/internal/stream-progress.js +195 -0
- package/dist/internal/tool-result.d.ts +20 -0
- package/dist/internal/tool-result.d.ts.map +1 -0
- package/dist/internal/tool-result.js +45 -0
- package/dist/internal/trajectory.d.ts +56 -0
- package/dist/internal/trajectory.d.ts.map +1 -0
- package/dist/internal/trajectory.js +248 -0
- package/dist/internal/workspace.d.ts +42 -0
- package/dist/internal/workspace.d.ts.map +1 -0
- package/dist/internal/workspace.js +199 -0
- package/dist/playground/assets/cursor-icons-outline-oY2V_mvK.woff2 +0 -0
- package/dist/playground/assets/index-1K-hG-7p.css +1 -0
- package/dist/playground/assets/index-FlWjhg3x.js +79 -0
- package/dist/playground/index.html +17 -0
- package/dist/reminders.d.ts +23 -0
- package/dist/reminders.d.ts.map +1 -0
- package/dist/reminders.js +22 -0
- package/dist/schedules.d.ts +23 -0
- package/dist/schedules.d.ts.map +1 -0
- package/dist/schedules.js +26 -0
- package/dist/skills.d.ts +22 -0
- package/dist/skills.d.ts.map +1 -0
- package/dist/skills.js +22 -0
- package/dist/tools.d.ts +92 -0
- package/dist/tools.d.ts.map +1 -0
- package/dist/tools.js +68 -0
- package/dist/types.d.ts +1609 -0
- package/dist/types.d.ts.map +1 -0
- package/dist/types.js +80 -0
- package/docs/README.md +174 -0
- package/docs/ab.md +355 -0
- package/docs/building-with-agents.md +129 -0
- package/docs/concepts.md +177 -0
- package/docs/deployment.md +218 -0
- package/docs/evals.md +383 -0
- package/docs/guides/agent-to-agent.md +118 -0
- package/docs/guides/cloud-runtime.md +122 -0
- package/docs/guides/github.md +172 -0
- package/docs/guides/human-in-the-loop.md +131 -0
- package/docs/guides/slack.md +245 -0
- package/docs/guides/webhooks.md +201 -0
- package/docs/hillclimbing.md +93 -0
- package/docs/quickstart.md +183 -0
- package/docs/reference/agent-config.md +154 -0
- package/docs/reference/channels.md +206 -0
- package/docs/reference/cli.md +257 -0
- package/docs/reference/connections.md +137 -0
- package/docs/reference/hooks.md +74 -0
- package/docs/reference/http-api.md +219 -0
- package/docs/reference/instructions.md +77 -0
- package/docs/reference/playground.md +67 -0
- package/docs/reference/project-layout.md +119 -0
- package/docs/reference/schedules.md +220 -0
- package/docs/reference/sessions.md +165 -0
- package/docs/reference/skills.md +79 -0
- package/docs/reference/subagents.md +74 -0
- package/docs/reference/tools.md +203 -0
- package/docs/scaffolding-agents.md +113 -0
- package/docs/troubleshooting.md +95 -0
- package/package.json +185 -0
- package/skills/ab/SKILL.md +139 -0
- package/skills/create-agent/SKILL.md +219 -0
- package/skills/debug/SKILL.md +71 -0
- package/skills/evals/SKILL.md +217 -0
- package/skills/framework-map/SKILL.md +142 -0
- package/skills/github/SKILL.md +144 -0
- package/skills/hillclimb/SKILL.md +162 -0
- package/skills/setup-slack/SKILL.md +180 -0
- package/src/ab.ts +675 -0
- package/src/bin/agent-serve.ts +949 -0
- package/src/channels/github/api.ts +374 -0
- package/src/channels/github/auth.ts +275 -0
- package/src/channels/github/binding.ts +223 -0
- package/src/channels/github/defaults.ts +111 -0
- package/src/channels/github/events.ts +72 -0
- package/src/channels/github/forward.ts +381 -0
- package/src/channels/github/github-channel.ts +2270 -0
- package/src/channels/github/index.ts +52 -0
- package/src/channels/github/replay.ts +387 -0
- package/src/channels/github/state.ts +95 -0
- package/src/channels/github/types.ts +490 -0
- package/src/channels/slack/api.ts +580 -0
- package/src/channels/slack/approvals.ts +198 -0
- package/src/channels/slack/assistant.ts +76 -0
- package/src/channels/slack/auth.ts +112 -0
- package/src/channels/slack/channel-watch.ts +219 -0
- package/src/channels/slack/constants.ts +30 -0
- package/src/channels/slack/credentials.ts +90 -0
- package/src/channels/slack/cursor-account.ts +166 -0
- package/src/channels/slack/defaults.ts +450 -0
- package/src/channels/slack/dispatch.ts +293 -0
- package/src/channels/slack/doctor.ts +135 -0
- package/src/channels/slack/event-id-dedupe.ts +30 -0
- package/src/channels/slack/inbound.ts +219 -0
- package/src/channels/slack/index.ts +45 -0
- package/src/channels/slack/init.ts +133 -0
- package/src/channels/slack/interactive.ts +194 -0
- package/src/channels/slack/links.ts +23 -0
- package/src/channels/slack/live-delivery.ts +201 -0
- package/src/channels/slack/log.ts +83 -0
- package/src/channels/slack/manifest.ts +179 -0
- package/src/channels/slack/pr-ref.ts +50 -0
- package/src/channels/slack/setup.ts +97 -0
- package/src/channels/slack/slack-channel.ts +374 -0
- package/src/channels/slack/socket-mode.ts +154 -0
- package/src/channels/slack/stream.ts +63 -0
- package/src/channels/slack/tool-approval-events.ts +136 -0
- package/src/channels/slack/types.ts +349 -0
- package/src/channels.ts +385 -0
- package/src/connections.ts +89 -0
- package/src/evals.ts +501 -0
- package/src/hooks.ts +33 -0
- package/src/index.ts +184 -0
- package/src/internal/ab-collector.ts +210 -0
- package/src/internal/ab-fold.ts +228 -0
- package/src/internal/ab-snapshot.ts +335 -0
- package/src/internal/approval-gate.ts +115 -0
- package/src/internal/auth.ts +39 -0
- package/src/internal/authored-alias-hooks.ts +69 -0
- package/src/internal/authored-loaders.ts +138 -0
- package/src/internal/bounded-int.ts +27 -0
- package/src/internal/brand.ts +29 -0
- package/src/internal/chat-client.ts +407 -0
- package/src/internal/cli-ax.ts +474 -0
- package/src/internal/cli-cursor.ts +275 -0
- package/src/internal/cli-github.ts +849 -0
- package/src/internal/cli-slack.ts +175 -0
- package/src/internal/cloud-merge.ts +41 -0
- package/src/internal/cron.ts +78 -0
- package/src/internal/cursor/account-mcp.ts +472 -0
- package/src/internal/cursor/backend-client.ts +614 -0
- package/src/internal/cursor/credentials.ts +401 -0
- package/src/internal/cursor-event-relay.ts +281 -0
- package/src/internal/cursor-relay-core.ts +444 -0
- package/src/internal/cursor-slack-relay.ts +242 -0
- package/src/internal/describe-error.ts +4 -0
- package/src/internal/dev-mode.ts +20 -0
- package/src/internal/discovery.ts +1709 -0
- package/src/internal/distribution.ts +97 -0
- package/src/internal/docs-site.ts +191 -0
- package/src/internal/duration.ts +39 -0
- package/src/internal/eval-run-store.ts +519 -0
- package/src/internal/eval-runner.ts +522 -0
- package/src/internal/event-log.ts +175 -0
- package/src/internal/event-mapper.ts +272 -0
- package/src/internal/frontmatter.ts +77 -0
- package/src/internal/github-fanout.ts +159 -0
- package/src/internal/handleAgentServeTrigger.ts +156 -0
- package/src/internal/host-platforms.ts +144 -0
- package/src/internal/http-channel.ts +546 -0
- package/src/internal/ids.ts +44 -0
- package/src/internal/init-project.ts +165 -0
- package/src/internal/json-dir-store.ts +109 -0
- package/src/internal/json-schema.ts +69 -0
- package/src/internal/log-ring.ts +59 -0
- package/src/internal/map-pool.ts +32 -0
- package/src/internal/mcp-endpoint.ts +518 -0
- package/src/internal/mcp-host.ts +211 -0
- package/src/internal/net.ts +31 -0
- package/src/internal/peer-connections.ts +100 -0
- package/src/internal/playground/static.ts +189 -0
- package/src/internal/playground/toolchain.ts +277 -0
- package/src/internal/reminder-runner.ts +756 -0
- package/src/internal/reminder-schedule.ts +106 -0
- package/src/internal/reminder-store.ts +145 -0
- package/src/internal/resolved-connections.ts +145 -0
- package/src/internal/review-comments.ts +268 -0
- package/src/internal/router.ts +145 -0
- package/src/internal/run-client.ts +320 -0
- package/src/internal/schedule-runner.ts +161 -0
- package/src/internal/sdk-runner.ts +315 -0
- package/src/internal/server.ts +1737 -0
- package/src/internal/session-engine.ts +1810 -0
- package/src/internal/session-store.ts +147 -0
- package/src/internal/session-urls.ts +40 -0
- package/src/internal/stream-progress.ts +254 -0
- package/src/internal/tool-result.ts +54 -0
- package/src/internal/trajectory.ts +306 -0
- package/src/internal/workspace.ts +223 -0
- package/src/reminders.ts +35 -0
- package/src/schedules.ts +37 -0
- package/src/skills.ts +27 -0
- package/src/tools.ts +127 -0
- package/src/types.ts +1848 -0
|
@@ -0,0 +1,522 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Discover and run filesystem evals under `<project>/evals/`.
|
|
3
|
+
*
|
|
4
|
+
* Case identity: single-`test` files use the path
|
|
5
|
+
* (`evals/weather/nyc.eval.ts` → `weather/nyc`); multi-`cases` files use
|
|
6
|
+
* `<fileId>/<case.id>` (`evals/weather.eval.ts` + `nyc` → `weather/nyc`).
|
|
7
|
+
* Drive with `t.send`, assert with `t.succeeded` / `t.calledTool` /
|
|
8
|
+
* `t.messageIncludes` / `t.check`.
|
|
9
|
+
*/
|
|
10
|
+
|
|
11
|
+
import type { Dirent } from "node:fs";
|
|
12
|
+
import { readdir } from "node:fs/promises";
|
|
13
|
+
import { join, relative, resolve } from "node:path";
|
|
14
|
+
import { pathToFileURL } from "node:url";
|
|
15
|
+
import type {
|
|
16
|
+
EvalAssertionResult,
|
|
17
|
+
EvalCaseDefinition,
|
|
18
|
+
EvalConfigFile,
|
|
19
|
+
EvalDefinition,
|
|
20
|
+
EvalSendOptions,
|
|
21
|
+
EvalTestContext,
|
|
22
|
+
EvalTurnResult,
|
|
23
|
+
Expectation,
|
|
24
|
+
} from "../evals.js";
|
|
25
|
+
import {
|
|
26
|
+
EVAL_MAX_CONCURRENCY_LIMIT,
|
|
27
|
+
expandEvalDefinition,
|
|
28
|
+
matchesEvalIdFilter,
|
|
29
|
+
normalizeEvalMaxConcurrency,
|
|
30
|
+
} from "../evals.js";
|
|
31
|
+
import type { SessionEvent } from "../types.js";
|
|
32
|
+
import { isDefinitionOfKind } from "./brand.js";
|
|
33
|
+
import { mapPool } from "./map-pool.js";
|
|
34
|
+
import { RunSessionTimeoutError, runSession } from "./run-client.js";
|
|
35
|
+
import { buildTrajectory } from "./trajectory.js";
|
|
36
|
+
|
|
37
|
+
export interface DiscoveredEval {
|
|
38
|
+
/** Runnable datapoint id (`weather/nyc` or file id for single-test files). */
|
|
39
|
+
id: string;
|
|
40
|
+
/** File id under `evals/` (`weather` for `evals/weather.eval.ts`). */
|
|
41
|
+
fileId: string;
|
|
42
|
+
path: string;
|
|
43
|
+
definition: EvalCaseDefinition;
|
|
44
|
+
}
|
|
45
|
+
|
|
46
|
+
export interface EvalRunResult {
|
|
47
|
+
id: string;
|
|
48
|
+
path: string;
|
|
49
|
+
description?: string;
|
|
50
|
+
ok: boolean;
|
|
51
|
+
assertions: EvalAssertionResult[];
|
|
52
|
+
sessionId?: string;
|
|
53
|
+
/** Messages passed to `t.send` (eval inputs). */
|
|
54
|
+
inputs: string[];
|
|
55
|
+
finalText?: string;
|
|
56
|
+
tools?: string[];
|
|
57
|
+
/** Tool calls with args/results from the trajectory. */
|
|
58
|
+
toolCalls?: Array<{
|
|
59
|
+
toolName: string;
|
|
60
|
+
args?: unknown;
|
|
61
|
+
output?: unknown;
|
|
62
|
+
isError: boolean;
|
|
63
|
+
}>;
|
|
64
|
+
error?: string;
|
|
65
|
+
logs: string[];
|
|
66
|
+
durationMs: number;
|
|
67
|
+
}
|
|
68
|
+
|
|
69
|
+
export interface EvalSelectionOptions {
|
|
70
|
+
filterIds?: string[];
|
|
71
|
+
tags?: string[];
|
|
72
|
+
}
|
|
73
|
+
|
|
74
|
+
export interface RunEvalsOptions extends EvalSelectionOptions {
|
|
75
|
+
projectRoot: string;
|
|
76
|
+
baseUrl: string;
|
|
77
|
+
timeoutMs?: number;
|
|
78
|
+
bearerToken?: string;
|
|
79
|
+
verbose?: boolean;
|
|
80
|
+
onEvent?: (event: SessionEvent) => void;
|
|
81
|
+
/** Fired when a case begins executing (after concurrency slot acquired). */
|
|
82
|
+
onCaseStart?: (id: string) => void;
|
|
83
|
+
/** Fired when a case finishes (pass or fail). */
|
|
84
|
+
onCaseDone?: (result: EvalRunResult) => void;
|
|
85
|
+
/** Fired for each session event during a case (for live progress / sessionId). */
|
|
86
|
+
onCaseEvent?: (id: string, event: SessionEvent) => void;
|
|
87
|
+
importModule?: (absolutePath: string) => Promise<Record<string, unknown>>;
|
|
88
|
+
/**
|
|
89
|
+
* Pre-discovered evals + config. When set, skips filesystem discovery and
|
|
90
|
+
* selection filters (caller already chose the case list).
|
|
91
|
+
*/
|
|
92
|
+
discovered?: {
|
|
93
|
+
evals: readonly DiscoveredEval[];
|
|
94
|
+
config: Partial<EvalConfigFile>;
|
|
95
|
+
};
|
|
96
|
+
}
|
|
97
|
+
|
|
98
|
+
async function defaultImport(
|
|
99
|
+
absolutePath: string
|
|
100
|
+
): Promise<Record<string, unknown>> {
|
|
101
|
+
return (await import(pathToFileURL(absolutePath).href)) as Record<
|
|
102
|
+
string,
|
|
103
|
+
unknown
|
|
104
|
+
>;
|
|
105
|
+
}
|
|
106
|
+
|
|
107
|
+
async function listEvalFiles(dir: string): Promise<string[]> {
|
|
108
|
+
const out: string[] = [];
|
|
109
|
+
async function walk(current: string): Promise<void> {
|
|
110
|
+
let entries: Dirent[];
|
|
111
|
+
try {
|
|
112
|
+
entries = await readdir(current, { withFileTypes: true });
|
|
113
|
+
} catch {
|
|
114
|
+
return;
|
|
115
|
+
}
|
|
116
|
+
for (const entry of entries) {
|
|
117
|
+
if (entry.name.startsWith(".")) {
|
|
118
|
+
continue;
|
|
119
|
+
}
|
|
120
|
+
const abs = join(current, entry.name);
|
|
121
|
+
if (entry.isDirectory()) {
|
|
122
|
+
if (entry.name === "node_modules") {
|
|
123
|
+
continue;
|
|
124
|
+
}
|
|
125
|
+
await walk(abs);
|
|
126
|
+
} else if (entry.isFile()) {
|
|
127
|
+
out.push(abs);
|
|
128
|
+
}
|
|
129
|
+
}
|
|
130
|
+
}
|
|
131
|
+
await walk(dir);
|
|
132
|
+
return out.sort();
|
|
133
|
+
}
|
|
134
|
+
|
|
135
|
+
export async function discoverEvals(
|
|
136
|
+
projectRoot: string,
|
|
137
|
+
importModule: (
|
|
138
|
+
absolutePath: string
|
|
139
|
+
) => Promise<Record<string, unknown>> = defaultImport
|
|
140
|
+
): Promise<{
|
|
141
|
+
evals: DiscoveredEval[];
|
|
142
|
+
config: Partial<EvalConfigFile>;
|
|
143
|
+
}> {
|
|
144
|
+
const root = resolve(projectRoot);
|
|
145
|
+
const evalsDir = join(root, "evals");
|
|
146
|
+
const files = await listEvalFiles(evalsDir);
|
|
147
|
+
const evals: DiscoveredEval[] = [];
|
|
148
|
+
let config: Partial<EvalConfigFile> = {};
|
|
149
|
+
|
|
150
|
+
for (const abs of files) {
|
|
151
|
+
const rel = relative(evalsDir, abs).replace(/\\/g, "/");
|
|
152
|
+
if (rel === "evals.config.ts" || rel === "evals.config.js") {
|
|
153
|
+
const mod = await importModule(abs);
|
|
154
|
+
if (mod.default !== null && typeof mod.default === "object") {
|
|
155
|
+
config = mod.default as Partial<EvalConfigFile>;
|
|
156
|
+
}
|
|
157
|
+
continue;
|
|
158
|
+
}
|
|
159
|
+
if (!rel.endsWith(".eval.ts") && !rel.endsWith(".eval.js")) {
|
|
160
|
+
continue;
|
|
161
|
+
}
|
|
162
|
+
const mod = await importModule(abs);
|
|
163
|
+
const def = mod.default;
|
|
164
|
+
if (!isDefinitionOfKind(def, "eval")) {
|
|
165
|
+
throw new Error(`${rel}: default export must be defineEval({ ... })`);
|
|
166
|
+
}
|
|
167
|
+
const fileId = rel.replace(/\.eval\.(ts|js)$/, "").replace(/\/index$/, "");
|
|
168
|
+
try {
|
|
169
|
+
for (const expanded of expandEvalDefinition(
|
|
170
|
+
fileId,
|
|
171
|
+
def as EvalDefinition
|
|
172
|
+
)) {
|
|
173
|
+
evals.push({
|
|
174
|
+
id: expanded.id,
|
|
175
|
+
fileId,
|
|
176
|
+
path: abs,
|
|
177
|
+
definition: expanded.definition,
|
|
178
|
+
});
|
|
179
|
+
}
|
|
180
|
+
} catch (error) {
|
|
181
|
+
const message = error instanceof Error ? error.message : String(error);
|
|
182
|
+
throw new Error(`${rel}: ${message}`);
|
|
183
|
+
}
|
|
184
|
+
}
|
|
185
|
+
|
|
186
|
+
return { evals, config };
|
|
187
|
+
}
|
|
188
|
+
|
|
189
|
+
/** Apply id/tag filters used by CLI, playground, and `runDiscoveredEvals`. */
|
|
190
|
+
export function filterDiscoveredEvals(
|
|
191
|
+
evals: readonly DiscoveredEval[],
|
|
192
|
+
options: EvalSelectionOptions = {}
|
|
193
|
+
): DiscoveredEval[] {
|
|
194
|
+
let selected = evals as DiscoveredEval[];
|
|
195
|
+
if (options.filterIds !== undefined && options.filterIds.length > 0) {
|
|
196
|
+
selected = selected.filter((e) =>
|
|
197
|
+
options.filterIds!.some((f) => matchesEvalIdFilter(e.id, f))
|
|
198
|
+
);
|
|
199
|
+
}
|
|
200
|
+
if (options.tags !== undefined && options.tags.length > 0) {
|
|
201
|
+
selected = selected.filter((e) =>
|
|
202
|
+
(e.definition.tags ?? []).some((t) => options.tags!.includes(t))
|
|
203
|
+
);
|
|
204
|
+
}
|
|
205
|
+
return selected;
|
|
206
|
+
}
|
|
207
|
+
|
|
208
|
+
export async function runDiscoveredEvals(
|
|
209
|
+
options: RunEvalsOptions
|
|
210
|
+
): Promise<EvalRunResult[]> {
|
|
211
|
+
let selected: readonly DiscoveredEval[];
|
|
212
|
+
let config: Partial<EvalConfigFile>;
|
|
213
|
+
|
|
214
|
+
if (options.discovered !== undefined) {
|
|
215
|
+
selected = options.discovered.evals;
|
|
216
|
+
config = options.discovered.config;
|
|
217
|
+
} else {
|
|
218
|
+
const importModule = options.importModule ?? defaultImport;
|
|
219
|
+
const discovered = await discoverEvals(options.projectRoot, importModule);
|
|
220
|
+
config = discovered.config;
|
|
221
|
+
selected = filterDiscoveredEvals(discovered.evals, {
|
|
222
|
+
filterIds: options.filterIds,
|
|
223
|
+
tags: options.tags,
|
|
224
|
+
});
|
|
225
|
+
}
|
|
226
|
+
|
|
227
|
+
if (selected.length === 0) {
|
|
228
|
+
return [];
|
|
229
|
+
}
|
|
230
|
+
|
|
231
|
+
const maxConcurrency = resolveEvalMaxConcurrency(config);
|
|
232
|
+
const defaultTimeoutMs = options.timeoutMs ?? config.timeoutMs ?? 180_000;
|
|
233
|
+
|
|
234
|
+
return mapPool(selected, maxConcurrency, async (discovered) => {
|
|
235
|
+
if (options.verbose) {
|
|
236
|
+
process.stderr.write(` eval ${discovered.id}…\n`);
|
|
237
|
+
}
|
|
238
|
+
options.onCaseStart?.(discovered.id);
|
|
239
|
+
const result = await runOneEval(discovered, {
|
|
240
|
+
baseUrl: options.baseUrl,
|
|
241
|
+
defaultTimeoutMs,
|
|
242
|
+
bearerToken: options.bearerToken,
|
|
243
|
+
verbose: Boolean(options.verbose),
|
|
244
|
+
onEvent: (event) => {
|
|
245
|
+
options.onCaseEvent?.(discovered.id, event);
|
|
246
|
+
options.onEvent?.(event);
|
|
247
|
+
},
|
|
248
|
+
});
|
|
249
|
+
options.onCaseDone?.(result);
|
|
250
|
+
return result;
|
|
251
|
+
});
|
|
252
|
+
}
|
|
253
|
+
|
|
254
|
+
function resolveEvalMaxConcurrency(config: Partial<EvalConfigFile>): number {
|
|
255
|
+
if (config.maxConcurrency === undefined) {
|
|
256
|
+
throw new Error(
|
|
257
|
+
`evals/evals.config.ts must define maxConcurrency (1–${EVAL_MAX_CONCURRENCY_LIMIT}; hard limit is due to model provider request limits). Example:\n` +
|
|
258
|
+
` import { defineEvalConfig } from "@anysphere/agent-serve/evals";\n` +
|
|
259
|
+
` export default defineEvalConfig({ maxConcurrency: 20 });`
|
|
260
|
+
);
|
|
261
|
+
}
|
|
262
|
+
return normalizeEvalMaxConcurrency(config.maxConcurrency);
|
|
263
|
+
}
|
|
264
|
+
|
|
265
|
+
async function runOneEval(
|
|
266
|
+
discovered: DiscoveredEval,
|
|
267
|
+
opts: {
|
|
268
|
+
baseUrl: string;
|
|
269
|
+
defaultTimeoutMs: number;
|
|
270
|
+
bearerToken?: string;
|
|
271
|
+
verbose: boolean;
|
|
272
|
+
onEvent?: (event: SessionEvent) => void;
|
|
273
|
+
}
|
|
274
|
+
): Promise<EvalRunResult> {
|
|
275
|
+
const logs: string[] = [];
|
|
276
|
+
const assertions: EvalAssertionResult[] = [];
|
|
277
|
+
const def = discovered.definition;
|
|
278
|
+
const timeoutMs = def.timeoutMs ?? opts.defaultTimeoutMs;
|
|
279
|
+
const startedAt = Date.now();
|
|
280
|
+
|
|
281
|
+
let sessionId: string | undefined;
|
|
282
|
+
let continuationToken: string | undefined;
|
|
283
|
+
const allEvents: SessionEvent[] = [];
|
|
284
|
+
const inputs: string[] = [];
|
|
285
|
+
let reply: string | null = null;
|
|
286
|
+
|
|
287
|
+
const record = (name: string, passed: boolean, detail?: string): void => {
|
|
288
|
+
assertions.push({
|
|
289
|
+
name,
|
|
290
|
+
passed,
|
|
291
|
+
...(detail === undefined ? {} : { detail }),
|
|
292
|
+
});
|
|
293
|
+
};
|
|
294
|
+
|
|
295
|
+
const toolsUsed = (): string[] =>
|
|
296
|
+
buildTrajectory(allEvents).turns.flatMap((t) =>
|
|
297
|
+
t.toolCalls.map((c) => c.toolName)
|
|
298
|
+
);
|
|
299
|
+
|
|
300
|
+
const toolCallIo = () =>
|
|
301
|
+
buildTrajectory(allEvents).turns.flatMap((turn) =>
|
|
302
|
+
turn.toolCalls.map((c) => ({
|
|
303
|
+
toolName: c.toolName,
|
|
304
|
+
...(c.args === undefined ? {} : { args: c.args }),
|
|
305
|
+
...(c.output === undefined ? {} : { output: c.output }),
|
|
306
|
+
isError: c.isError,
|
|
307
|
+
}))
|
|
308
|
+
);
|
|
309
|
+
|
|
310
|
+
const t: EvalTestContext = {
|
|
311
|
+
get reply() {
|
|
312
|
+
return reply;
|
|
313
|
+
},
|
|
314
|
+
get sessionId() {
|
|
315
|
+
return sessionId;
|
|
316
|
+
},
|
|
317
|
+
get events() {
|
|
318
|
+
return allEvents;
|
|
319
|
+
},
|
|
320
|
+
async send(
|
|
321
|
+
message: string,
|
|
322
|
+
options: EvalSendOptions = {}
|
|
323
|
+
): Promise<EvalTurnResult> {
|
|
324
|
+
inputs.push(message);
|
|
325
|
+
const result = await runSession({
|
|
326
|
+
baseUrl: opts.baseUrl,
|
|
327
|
+
messages: [message],
|
|
328
|
+
sessionId,
|
|
329
|
+
continuationToken,
|
|
330
|
+
// Keep defineEval traffic out of live defineAB enrollment/metrics.
|
|
331
|
+
purpose: "eval",
|
|
332
|
+
bearerToken: opts.bearerToken,
|
|
333
|
+
workspaceDir: options.workspaceDir,
|
|
334
|
+
workspaceFiles: options.workspaceFiles,
|
|
335
|
+
timeoutMs,
|
|
336
|
+
onEvent: opts.onEvent,
|
|
337
|
+
});
|
|
338
|
+
sessionId = result.sessionId;
|
|
339
|
+
continuationToken = result.continuationToken;
|
|
340
|
+
allEvents.push(...result.events);
|
|
341
|
+
if (result.trajectory.finalText !== "") {
|
|
342
|
+
reply = result.trajectory.finalText;
|
|
343
|
+
}
|
|
344
|
+
return {
|
|
345
|
+
message: result.trajectory.finalText,
|
|
346
|
+
sessionId: result.sessionId,
|
|
347
|
+
events: result.events,
|
|
348
|
+
toolCalls: result.trajectory.turns.flatMap((turn) =>
|
|
349
|
+
turn.toolCalls.map((c) => c.toolName)
|
|
350
|
+
),
|
|
351
|
+
ok: result.ok,
|
|
352
|
+
};
|
|
353
|
+
},
|
|
354
|
+
succeeded() {
|
|
355
|
+
const traj = buildTrajectory(allEvents);
|
|
356
|
+
record(
|
|
357
|
+
"succeeded",
|
|
358
|
+
traj.ok && traj.turns.length > 0,
|
|
359
|
+
traj.ok ? undefined : "turn failed"
|
|
360
|
+
);
|
|
361
|
+
},
|
|
362
|
+
calledTool(name: string) {
|
|
363
|
+
const tools = toolsUsed();
|
|
364
|
+
record(
|
|
365
|
+
`calledTool(${name})`,
|
|
366
|
+
tools.includes(name),
|
|
367
|
+
tools.includes(name) ? undefined : `got [${tools.join(", ")}]`
|
|
368
|
+
);
|
|
369
|
+
},
|
|
370
|
+
notCalledTool(name: string) {
|
|
371
|
+
const tools = toolsUsed();
|
|
372
|
+
record(
|
|
373
|
+
`notCalledTool(${name})`,
|
|
374
|
+
!tools.includes(name),
|
|
375
|
+
tools.includes(name) ? "tool was called" : undefined
|
|
376
|
+
);
|
|
377
|
+
},
|
|
378
|
+
messageIncludes(token: string | RegExp) {
|
|
379
|
+
const text = reply ?? "";
|
|
380
|
+
const passed =
|
|
381
|
+
typeof token === "string" ? text.includes(token) : token.test(text);
|
|
382
|
+
record(
|
|
383
|
+
`messageIncludes(${String(token)})`,
|
|
384
|
+
passed,
|
|
385
|
+
passed ? undefined : `reply=${JSON.stringify(text.slice(0, 120))}`
|
|
386
|
+
);
|
|
387
|
+
},
|
|
388
|
+
check(value: unknown, assertion: Expectation) {
|
|
389
|
+
const { passed, detail } = matchExpectation(value, assertion);
|
|
390
|
+
record(`check(${assertion.kind})`, passed, detail);
|
|
391
|
+
},
|
|
392
|
+
log(message: string) {
|
|
393
|
+
logs.push(message);
|
|
394
|
+
if (opts.verbose) {
|
|
395
|
+
process.stderr.write(` ${message}\n`);
|
|
396
|
+
}
|
|
397
|
+
},
|
|
398
|
+
};
|
|
399
|
+
|
|
400
|
+
try {
|
|
401
|
+
await def.test(t);
|
|
402
|
+
} catch (error) {
|
|
403
|
+
const message = error instanceof Error ? error.message : String(error);
|
|
404
|
+
if (error instanceof RunSessionTimeoutError && sessionId === undefined) {
|
|
405
|
+
sessionId = error.sessionId;
|
|
406
|
+
}
|
|
407
|
+
return {
|
|
408
|
+
id: discovered.id,
|
|
409
|
+
path: discovered.path,
|
|
410
|
+
...(def.description === undefined
|
|
411
|
+
? {}
|
|
412
|
+
: { description: def.description }),
|
|
413
|
+
ok: false,
|
|
414
|
+
assertions,
|
|
415
|
+
error: message,
|
|
416
|
+
logs,
|
|
417
|
+
inputs,
|
|
418
|
+
toolCalls: toolCallIo(),
|
|
419
|
+
...(sessionId === undefined ? {} : { sessionId }),
|
|
420
|
+
...(reply === null ? {} : { finalText: reply }),
|
|
421
|
+
tools: toolsUsed(),
|
|
422
|
+
durationMs: Date.now() - startedAt,
|
|
423
|
+
};
|
|
424
|
+
}
|
|
425
|
+
|
|
426
|
+
const traj = buildTrajectory(allEvents);
|
|
427
|
+
const ok =
|
|
428
|
+
assertions.length > 0
|
|
429
|
+
? assertions.every((a) => a.passed)
|
|
430
|
+
: traj.ok && traj.turns.length > 0;
|
|
431
|
+
|
|
432
|
+
return {
|
|
433
|
+
id: discovered.id,
|
|
434
|
+
path: discovered.path,
|
|
435
|
+
...(def.description === undefined ? {} : { description: def.description }),
|
|
436
|
+
ok,
|
|
437
|
+
assertions,
|
|
438
|
+
logs,
|
|
439
|
+
inputs,
|
|
440
|
+
toolCalls: toolCallIo(),
|
|
441
|
+
...(sessionId === undefined ? {} : { sessionId }),
|
|
442
|
+
finalText: reply ?? traj.finalText,
|
|
443
|
+
tools: toolsUsed(),
|
|
444
|
+
durationMs: Date.now() - startedAt,
|
|
445
|
+
};
|
|
446
|
+
}
|
|
447
|
+
|
|
448
|
+
function matchExpectation(
|
|
449
|
+
value: unknown,
|
|
450
|
+
assertion: Expectation
|
|
451
|
+
): { passed: boolean; detail?: string } {
|
|
452
|
+
switch (assertion.kind) {
|
|
453
|
+
case "includes": {
|
|
454
|
+
const text = value === null || value === undefined ? "" : String(value);
|
|
455
|
+
const passed =
|
|
456
|
+
typeof assertion.value === "string"
|
|
457
|
+
? text.includes(assertion.value)
|
|
458
|
+
: assertion.value.test(text);
|
|
459
|
+
return {
|
|
460
|
+
passed,
|
|
461
|
+
...(passed
|
|
462
|
+
? {}
|
|
463
|
+
: {
|
|
464
|
+
detail: `value=${JSON.stringify(text.slice(0, 120))}`,
|
|
465
|
+
}),
|
|
466
|
+
};
|
|
467
|
+
}
|
|
468
|
+
case "equals": {
|
|
469
|
+
const passed = deepEqual(value, assertion.value);
|
|
470
|
+
return {
|
|
471
|
+
passed,
|
|
472
|
+
...(passed ? {} : { detail: `got ${JSON.stringify(value)}` }),
|
|
473
|
+
};
|
|
474
|
+
}
|
|
475
|
+
case "satisfies": {
|
|
476
|
+
const passed = assertion.predicate(value);
|
|
477
|
+
return {
|
|
478
|
+
passed,
|
|
479
|
+
...(passed ? {} : { detail: assertion.label }),
|
|
480
|
+
};
|
|
481
|
+
}
|
|
482
|
+
}
|
|
483
|
+
}
|
|
484
|
+
|
|
485
|
+
function deepEqual(a: unknown, b: unknown): boolean {
|
|
486
|
+
if (Object.is(a, b)) {
|
|
487
|
+
return true;
|
|
488
|
+
}
|
|
489
|
+
if (
|
|
490
|
+
typeof a !== "object" ||
|
|
491
|
+
typeof b !== "object" ||
|
|
492
|
+
a === null ||
|
|
493
|
+
b === null
|
|
494
|
+
) {
|
|
495
|
+
return false;
|
|
496
|
+
}
|
|
497
|
+
if (Array.isArray(a) !== Array.isArray(b)) {
|
|
498
|
+
return false;
|
|
499
|
+
}
|
|
500
|
+
if (Array.isArray(a) && Array.isArray(b)) {
|
|
501
|
+
if (a.length !== b.length) {
|
|
502
|
+
return false;
|
|
503
|
+
}
|
|
504
|
+
return a.every((item, i) => deepEqual(item, b[i]));
|
|
505
|
+
}
|
|
506
|
+
const aKeys = Object.keys(a as object);
|
|
507
|
+
const bKeys = Object.keys(b as object);
|
|
508
|
+
if (aKeys.length !== bKeys.length) {
|
|
509
|
+
return false;
|
|
510
|
+
}
|
|
511
|
+
for (const key of aKeys) {
|
|
512
|
+
if (
|
|
513
|
+
!deepEqual(
|
|
514
|
+
(a as Record<string, unknown>)[key],
|
|
515
|
+
(b as Record<string, unknown>)[key]
|
|
516
|
+
)
|
|
517
|
+
) {
|
|
518
|
+
return false;
|
|
519
|
+
}
|
|
520
|
+
}
|
|
521
|
+
return true;
|
|
522
|
+
}
|
|
@@ -0,0 +1,175 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Durable, replayable per-session event log. Events are held in memory for
|
|
3
|
+
* live fan-out and appended to `events.ndjson` for durability; on process
|
|
4
|
+
* restart the file is hydrated back so streams can replay from any
|
|
5
|
+
* `startIndex`.
|
|
6
|
+
*/
|
|
7
|
+
|
|
8
|
+
import { appendFile, mkdir, readFile } from "node:fs/promises";
|
|
9
|
+
import { dirname } from "node:path";
|
|
10
|
+
import type { SessionEvent, SessionEventPayload } from "../types.js";
|
|
11
|
+
|
|
12
|
+
export type AppendableEvent = SessionEventPayload & { turnId?: string };
|
|
13
|
+
|
|
14
|
+
export class SessionEventLog {
|
|
15
|
+
private readonly events: SessionEvent[] = [];
|
|
16
|
+
private readonly listeners = new Set<(event: SessionEvent) => void>();
|
|
17
|
+
private persistChain: Promise<void> = Promise.resolve();
|
|
18
|
+
|
|
19
|
+
private constructor(
|
|
20
|
+
private readonly sessionId: string,
|
|
21
|
+
private readonly filePath: string,
|
|
22
|
+
private readonly onPersistError: (error: unknown) => void
|
|
23
|
+
) {}
|
|
24
|
+
|
|
25
|
+
static async open(args: {
|
|
26
|
+
sessionId: string;
|
|
27
|
+
filePath: string;
|
|
28
|
+
onPersistError?: (error: unknown) => void;
|
|
29
|
+
}): Promise<SessionEventLog> {
|
|
30
|
+
const log = new SessionEventLog(
|
|
31
|
+
args.sessionId,
|
|
32
|
+
args.filePath,
|
|
33
|
+
args.onPersistError ?? (() => {})
|
|
34
|
+
);
|
|
35
|
+
await log.hydrate();
|
|
36
|
+
return log;
|
|
37
|
+
}
|
|
38
|
+
|
|
39
|
+
private async hydrate(): Promise<void> {
|
|
40
|
+
let raw: string;
|
|
41
|
+
try {
|
|
42
|
+
raw = await readFile(this.filePath, "utf8");
|
|
43
|
+
} catch {
|
|
44
|
+
return;
|
|
45
|
+
}
|
|
46
|
+
for (const line of raw.split("\n")) {
|
|
47
|
+
if (line.trim() === "") {
|
|
48
|
+
continue;
|
|
49
|
+
}
|
|
50
|
+
try {
|
|
51
|
+
this.events.push(JSON.parse(line) as SessionEvent);
|
|
52
|
+
} catch {
|
|
53
|
+
// Skip torn trailing writes from a crashed process.
|
|
54
|
+
}
|
|
55
|
+
}
|
|
56
|
+
}
|
|
57
|
+
|
|
58
|
+
get count(): number {
|
|
59
|
+
return this.events.length;
|
|
60
|
+
}
|
|
61
|
+
|
|
62
|
+
/**
|
|
63
|
+
* Assign the next index, record the event, and fan out to live
|
|
64
|
+
* subscribers. Ordering is synchronous; file persistence is chained in
|
|
65
|
+
* the background and awaited by {@link flush}.
|
|
66
|
+
*/
|
|
67
|
+
append(payload: AppendableEvent): SessionEvent {
|
|
68
|
+
const { turnId, ...rest } = payload;
|
|
69
|
+
const event = {
|
|
70
|
+
index: this.events.length,
|
|
71
|
+
sessionId: this.sessionId,
|
|
72
|
+
...(turnId === undefined ? {} : { turnId }),
|
|
73
|
+
at: new Date().toISOString(),
|
|
74
|
+
...rest,
|
|
75
|
+
} as SessionEvent;
|
|
76
|
+
this.events.push(event);
|
|
77
|
+
for (const listener of [...this.listeners]) {
|
|
78
|
+
try {
|
|
79
|
+
listener(event);
|
|
80
|
+
} catch {
|
|
81
|
+
// Subscriber errors must not break the log.
|
|
82
|
+
}
|
|
83
|
+
}
|
|
84
|
+
const line = `${JSON.stringify(event)}\n`;
|
|
85
|
+
this.persistChain = this.persistChain
|
|
86
|
+
.then(() => appendFile(this.filePath, line, "utf8"))
|
|
87
|
+
.catch((error) => this.onPersistError(error));
|
|
88
|
+
return event;
|
|
89
|
+
}
|
|
90
|
+
|
|
91
|
+
/** Wait for pending file writes (used on close). */
|
|
92
|
+
async flush(): Promise<void> {
|
|
93
|
+
await this.persistChain;
|
|
94
|
+
}
|
|
95
|
+
|
|
96
|
+
snapshot(startIndex = 0): SessionEvent[] {
|
|
97
|
+
return startIndex <= 0 ? [...this.events] : this.events.slice(startIndex);
|
|
98
|
+
}
|
|
99
|
+
|
|
100
|
+
subscribe(listener: (event: SessionEvent) => void): () => void {
|
|
101
|
+
this.listeners.add(listener);
|
|
102
|
+
return () => {
|
|
103
|
+
this.listeners.delete(listener);
|
|
104
|
+
};
|
|
105
|
+
}
|
|
106
|
+
|
|
107
|
+
/**
|
|
108
|
+
* NDJSON stream: replays recorded events from `startIndex`, then follows
|
|
109
|
+
* live events until the consumer cancels.
|
|
110
|
+
*/
|
|
111
|
+
stream(startIndex = 0): ReadableStream<Uint8Array> {
|
|
112
|
+
const encoder = new TextEncoder();
|
|
113
|
+
let unsubscribe: (() => void) | undefined;
|
|
114
|
+
return new ReadableStream<Uint8Array>({
|
|
115
|
+
start: (controller) => {
|
|
116
|
+
const push = (event: SessionEvent): void => {
|
|
117
|
+
try {
|
|
118
|
+
controller.enqueue(encoder.encode(`${JSON.stringify(event)}\n`));
|
|
119
|
+
} catch {
|
|
120
|
+
unsubscribe?.();
|
|
121
|
+
}
|
|
122
|
+
};
|
|
123
|
+
// Synchronous snapshot + subscribe: no events can interleave.
|
|
124
|
+
for (const event of this.snapshot(startIndex)) {
|
|
125
|
+
push(event);
|
|
126
|
+
}
|
|
127
|
+
unsubscribe = this.subscribe((event) => {
|
|
128
|
+
if (event.index >= startIndex) {
|
|
129
|
+
push(event);
|
|
130
|
+
}
|
|
131
|
+
});
|
|
132
|
+
},
|
|
133
|
+
cancel: () => {
|
|
134
|
+
unsubscribe?.();
|
|
135
|
+
},
|
|
136
|
+
});
|
|
137
|
+
}
|
|
138
|
+
}
|
|
139
|
+
|
|
140
|
+
/** Lazily opens one {@link SessionEventLog} per session. */
|
|
141
|
+
export class EventLogStore {
|
|
142
|
+
private readonly logs = new Map<string, Promise<SessionEventLog>>();
|
|
143
|
+
|
|
144
|
+
constructor(
|
|
145
|
+
private readonly eventFilePathFor: (sessionId: string) => string,
|
|
146
|
+
private readonly onPersistError: (
|
|
147
|
+
sessionId: string,
|
|
148
|
+
error: unknown
|
|
149
|
+
) => void = () => {}
|
|
150
|
+
) {}
|
|
151
|
+
|
|
152
|
+
get(sessionId: string): Promise<SessionEventLog> {
|
|
153
|
+
let log = this.logs.get(sessionId);
|
|
154
|
+
if (log === undefined) {
|
|
155
|
+
const filePath = this.eventFilePathFor(sessionId);
|
|
156
|
+
log = mkdir(dirname(filePath), { recursive: true }).then(() =>
|
|
157
|
+
SessionEventLog.open({
|
|
158
|
+
sessionId,
|
|
159
|
+
filePath,
|
|
160
|
+
onPersistError: (error) => this.onPersistError(sessionId, error),
|
|
161
|
+
})
|
|
162
|
+
);
|
|
163
|
+
this.logs.set(sessionId, log);
|
|
164
|
+
}
|
|
165
|
+
return log;
|
|
166
|
+
}
|
|
167
|
+
|
|
168
|
+
async flushAll(): Promise<void> {
|
|
169
|
+
await Promise.all(
|
|
170
|
+
[...this.logs.values()].map((logPromise) =>
|
|
171
|
+
logPromise.then((log) => log.flush()).catch(() => {})
|
|
172
|
+
)
|
|
173
|
+
);
|
|
174
|
+
}
|
|
175
|
+
}
|