feat: A2A v2 research, architecture design & doc fixes (#31, #33, #34)#36
feat: A2A v2 research, architecture design & doc fixes (#31, #33, #34)#36
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 7e825263db
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
docs/KNOWN_ISSUES.md
Outdated
| ```yaml | ||
| channels: | ||
| feishu: | ||
| groupSessionScope: “group_topic” |
There was a problem hiding this comment.
The YAML snippet here uses typographic quotes (“group_topic”) instead of plain ASCII quotes, so users who copy this config will end up with invalid YAML (or a value that OpenClaw cannot parse as intended) when enabling groupSessionScope. Because this section is the documented resolution path for the Feishu isolation issue, the snippet should use standard "group_topic" to avoid setup failures.
Useful? React with 👍 / 👎.
…, #34) Address user-reported issues and integrate new platform capabilities: Discord (Issue #34 — routing bug): - Add Step 5b: channel permission isolation to prevent agent conversation mixing (single-bot and multi-bot approaches) - Update multi-account support status (available in current versions) Feishu (Issue #31 — multi-agent support): - Document groupSessionScope: "group_topic" for per-topic session isolation (OpenClaw >= 2026.3.1), shown as incremental config patch - Add CLI alternative: openclaw config set - Add SecretRef bug caveat (openclaw #47436) in multi-app section - Fix issue citation: #10242 (DM threads) → #29791 (group chat topics) Getting Started (Issue #33 — beginner experience): - Fix deployment order with clear "Manual" vs "Agent-automated" labels - Note that steps 5-9 can be fully automated via DEPLOY.md - Add prerequisites checklist covering all platforms - Add common mistakes section (5 items) Known Issues: - Add Feishu P1 (resolved) entry with groupSessionScope resolution - Fix typographic quotes in YAML examples (would cause parse failures) All changes in both Chinese and English versions. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
AlexAnys
force-pushed
the
feat/a2a-v2-research-and-doc-fixes
branch
from
61c23c0 to
6e82fe8
Compare
1 participant
Summary
Multi-agent harness research and architecture design for evolving OpenCrew's A2A from "delegation-only" to "true multi-agent collaboration," plus documentation fixes for issues #31, #33, and #34.
Documentation Fixes (Commit 1 — user-facing)
groupSessionScope: "group_topic"for per-topic session isolationA2A v2 Research & Architecture (Commit 2 — harness artifacts)
Phase 1 — Research (3 parallel agents):
allowBots + requireMention + multi-account. Self-loop filter is per-bot-user-ID → naturally bypassed.groupSessionScope: "group_topic"solves P0 session sharing. But bot messages DON'T trigger other bots (platform limitation) → sessions_send still required.Phase 2 — Architecture (2 parallel agents):
Phase 3 — QA Verification:
Key Architectural Insight
The "true A2A" value from harness design methodology:
Files
Test plan