v0.11.1
What's Changed
- feat(agents): reliability foundation for long tasks by @andrew-ge-wu in #129
- feat(agents): durable task memory — pinned goal + fact graduation by @andrew-ge-wu in #130
- chore(providers): trip the SSE parser when N consecutive parses fail by @andrew-ge-wu in #131
- feat(agents): warm-summary retry + degraded compaction status by @andrew-ge-wu in #132
- feat(tools): central dispatch + ToolResult union + external zod by @andrew-ge-wu in #133
- feat(tools): wrap registered tools so the agent loop also flows through dispatch by @andrew-ge-wu in #134
- fix(tools): return error envelopes instead of throwing on transport/d… by @andrew-ge-wu in #135
- feat(agents): surface tool error codes as first-class chunk fields by @andrew-ge-wu in #136
- feat(tools): standard error-code vocabulary across HTTP + fs tools by @andrew-ge-wu in #137
- feat(agents): provider error classifiers + transient auto-retry by @andrew-ge-wu in #138
- feat(api): client request() timeout + transient-status retry by @andrew-ge-wu in #139
- feat(db): SQLITE_BUSY retry, constraint translation, migration safety, error code by @andrew-ge-wu in #140
- feat(chat): code-aware error card consuming the unified error vocabulary by @andrew-ge-wu in #141
- chore: dedupe estimateTokens + inline thin modules + drop dead script by @andrew-ge-wu in #142
- refactor(tools): split atlassian.ts (3130 LOC) into focused sub-modules by @andrew-ge-wu in #143
- feat(tools): self-correcting hints + slim system-prompt playbook by @andrew-ge-wu in #144
- feat(agents): output-validator hit-rate telemetry + escape-hatch flag by @andrew-ge-wu in #145
- feat(logs): live server-log panel via console patch + SSE by @andrew-ge-wu in #146
- fix(chat): soften 8s SSE connect-timeout + Reconnect for stream drops by @andrew-ge-wu in #147
- feat(env): runtime overrides UI + restart endpoint + agent tools by @andrew-ge-wu in #148
- chore(release): 0.11.0 by @andrew-ge-wu in #149
- chore(release): 0.11.1 by @andrew-ge-wu in #150
Full Changelog: v0.10.0...v0.11.1