Commit Graph

19 Commits

Author SHA1 Message Date
lxowalle 2992eccbf0 feat: add request-scoped context policies (#2914)
* feat: add request-scoped context policies

Add named turn profiles under agents.defaults so callers can opt into
per-request context and tool policies without changing default chat behavior.

Profiles can disable history, system context, skill prompts, or tools, and can
limit skills/tools with allow lists. Wire profile selection through Pico message
payloads, agent turn execution, Web chat selection, and Web visual config.

Reject invalid turn profiles before saving config through Web APIs and document
the new request context policy behavior.

* fix: address turn profile review blockers

* feat: simplify request context policy config

* fix: suppress tool prompt when turn tools are disabled

* fix: enforce turn profile tool restrictions
2026-05-22 10:06:40 +08:00
lxowalle e7e21df354 fix(agent): honor explicit thinking off (#2898)
* fix(agent): honor explicit thinking off

* fix(agent): address thinking off lint failures

* Clarify unset thinking level display

* fix ci
2026-05-21 11:07:39 +08:00
LC b7db059544 feat(chat,seahorse): persist and display model_name across history (#2897)
* feat(chat,seahorse): persist and display model_name across history

* test(seahorse): fix lint regressions in repair coverage

* fix(pico): preserve model_name in live updates

* fix(pico): preserve model_name through live stream wrappers
2026-05-20 13:42:21 +08:00
lxowalle 639b32703a feat: support streaming (#2892)
* Support streaming

* fix: stream pico reasoning updates

Route Pico reasoning through the active streamer and hide empty thought placeholders.

* fix: harden configured streaming delivery

* fix ci

* fix split issue
2026-05-19 16:38:47 +08:00
lc6464 6ae7dc38b9 fix(openai_compat): align MiMo reasoning replay with DeepSeek 2026-05-14 20:59:01 +08:00
lxowalle b3a7b7ad64 feat: agent self evolution (#2847)
* feat: add agent self-evolution

* fix ci

* delete unused doc

* fix lint

* fix evolution review issues
2026-05-11 16:13:27 +08:00
Hoshina 057683d94c fix(agent): use runtime event kind for LLM retry 2026-05-04 15:06:34 +08:00
Mauro be67aed4dc Merge pull request #2677 from alexhoshina/feat/runtime-events-plan
Feat/runtime events
2026-05-03 23:15:25 +02:00
Mauro 272dee3fca Merge pull request #2669 from david1gp/fix/network-error-retry
feat(agent): add network error retry with configurable max retries and backoff
2026-05-03 20:18:18 +02:00
Hoshina 795ee362ea refactor(events): emit agent runtime events directly
Remove the legacy EventKind/Event envelope mapping and let agent event emission build pkg/events.Event values directly.

Keep HookMeta as the shared hook metadata shape and preserve legacy observe string aliases by mapping them to runtime event kinds.

Validation: GOCACHE=/tmp/picoclaw-go-cache go test ./pkg/agent; make lint
2026-04-26 16:55:02 +08:00
lc6464 6d04d15ce0 fix(tool-feedback): dedupe duplicate content and keep full explanations 2026-04-26 00:40:55 +08:00
lc6464 5cd10b594a feat(pico): add support for tool_calls in chat messages 2026-04-25 23:43:10 +08:00
David Siewert 06fad95719 feat(agent): add network error retry with configurable max retries and backoff
- Add isNetworkError detection for connection reset, broken pipe, read/write tcp, EOF
- Add retry logic with configurable exponential backoff for network errors
- Add config options max_llm_retries and llm_retry_backoff_secs in agents.defaults
- Network errors now retry with backoff (was previously not retried)
- Timeout errors now use configurable backoff instead of hardcoded 5s
- Default: 2 retries with 2s backoff (3 total attempts)
2026-04-25 19:08:46 +06:00
美電球 41f4d95597 Merge pull request #2657 from lc6464/fix-deepseek-v4-thinking-history
fix(reasoning): persist canonical history for DeepSeek and web chat
2026-04-25 15:08:48 +08:00
lc6464 bb0f983708 fix(reasoning): persist canonical history for DeepSeek and web chat 2026-04-24 21:45:41 +08:00
Hoshina 2e65b1be83 feat(agent): add structured prompt layering 2026-04-24 18:14:28 +08:00
wenjie cac4f21746 fix(tools): improve web search provider fallback (#2629)
- centralize web search provider readiness and resolution logic
- fall back when the configured provider is unavailable or invalid
- allow native-search-capable models to use built-in search without the client tool
- simplify the tools page and add direct access to web search settings
- add backend, agent, and integration tests for the new selection behavior
2026-04-23 15:39:16 +08:00
lxowalle 451db2f5d8 Feat(channels): unify animated tool feedback across chat channels and Pico (#2622)
* feat(channels): unify tool feedback animation across discord telegram and feishu

* fix(tool-feedback): unify fallback and single-message delivery

* fix(channels): finalize tool feedback in place

* fix ci

* feat: improve tool feedback

* fix review blockers in pico token cache and tool feedback

fix(provider): preserve function thought signatures

fix(feishu): recover tool feedback after edit fallback

* * delete dead code

* fix(pico): clean up tool feedback progress state

* fix ci

* fix(web): preserve tool feedback line breaks in chat

* fix(channels): preserve tool feedback progress state

fix(pico): preserve context usage when finalizing tool feedback

chore: record branch review pass

fix: preserve tool feedback finalization state

fix(web): handle pico history update fallback

* fix ci
2026-04-23 10:35:50 +08:00
sky5454 329e68e017 refactor(agent): Agent Looper refactor phase2, restructure pipeline and rename loop files to agent (#2585)
* refactor(agent): introduce interfaces for MessageBus and ChannelManager

Phase 2 of loop.go refactor — dependency inversion using adapter pattern.

- Add interfaces.MessageBus and interfaces.ChannelManager interfaces
- Create adapters/messagebus.go wrapping *bus.MessageBus
- Create adapters/channelmanager.go wrapping *channels.Manager
- Update AgentLoop to use interfaces instead of concrete types
- Update registerSharedTools to accept interfaces.MessageBus

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* refactor(agent): restructure pipeline and rename loop files

Pipeline refactoring:
- Split pipeline.go (1400 lines) into focused files:
  - pipeline_setup.go (~115 lines): SetupTurn method
  - pipeline_llm.go (~519 lines): CallLLM method
  - pipeline_execute.go (~693 lines): ExecuteTools method
  - pipeline_finalize.go (~78 lines): Finalize method
- Pipeline struct and NewPipeline remain in pipeline.go (~39 lines)

Agent file renaming:
- Rename loop_*.go to agent_*.go for consistent naming:
  - loop.go -> agent.go, loop_message.go -> agent_message.go, etc.
- Merge turn.go + turn_exec.go into turn_state.go
- Rename loop_turn.go -> turn_coord.go

Documentation:
- Update docs/pipeline-restructuring-plan.md
- Add docs/agent-rename-plan.md

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* fix(agent): code format  fixed

* refactor(agent): code test file added/renamed

* docs(agent): update agent refactor docs

* fix(agent): fix agent hardAbortX

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-04-21 10:55:50 +08:00