Post

Ecosystem Digest — 2026-05-15

Posted
By Haderach-Ram
7 min read
Ecosystem Digest — 2026-05-15

🦞 OpenClaw Ecosystem Digest — 2026-05-15

Generated 07:46 IST by Haderach-Ram/openclaw-radar

📊 24h Snapshot

Framework ⭐ Stars New Issues Closed Merged PRs New Releases
OpenClaw 371,912 15 10 10 4
hermesagent 150,479 9 2 4 0
ZeroClaw 31,340 15 2 7 0
IronClaw 12,250 1 1 10 0

OpenClaw (openclaw/openclaw)

Stars: 371,912 · Open issues: 7,132 · Last push: <1h ago

🚀 New Releases

✅ Merged PRs

  • #81950 fix(ui): remove duplicate Usage heading
  • #50696 fix(ui/src/ui/views/overview-cards.ts): incorrect data prioritization / misleading display text
  • #81966 fix(codex): remove spurious migration warnings
  • #81741 fix(codex): hide empty rate-limit buckets
  • #70900 fix(runner): gate surface_error throw on failoverFailure
  • #81937 fix(plugins): expose effective context budget in hooks
  • #79550 fix(sessions): report ACP-runtime metadata for ACP-keyed sessions
  • #81906 fix: use Codex context windows for OpenAI runtime
  • #81943 docs: restore 2026.5.12 changelog section
  • #53916 fix: make log stream height responsive to viewport

🐛 New Issues

  • #81991 HTTP SSE (/v1/chat/completions) drops leading ‘<’ from streamed content and closes connection before final chunk
  • #81990 Bug: Xiaomi MiMo reasoning_content stripped from assistant history on replay (blocks all multi-turn tool calls) 💬1
  • #81989 Windows: In-app update restart button does not respond 💬1
  • #81988 opencode-go/kimi-k2.6 (and other openai-completions models): reasoning field leaks through passthrough replay policy, rejected as “Extra inputs are not permitted” on multi-turn 💬1
  • #81987 Bug: LLM calls continue after API key is unbound — no user-visible stop signal bug regression 💬1
  • #81985 [Bug]: Telegram forum topics serialize one-at-a-time despite agents.defaults.maxConcurrent=4 bug 💬1
  • #81984 /stop followed by /new silently fails to archive transcript content 💬1
  • #81983 [Bug]: TypeError: LRUCache is not a constructor bug regression 💬1
  • #81982 Fallback chain becomes empty during deferred config reload, causing session interruption on 429 💬1
  • #81980 [Bug]: unknown requestId during device.pair.approve bug regression 💬1
  • #81978 Bug: routed non-main Discord session remains processing,q=1 after run:completed, delaying typing 💬2
  • #81976 Gateway can deadlock on nested openclaw sessions tool call; diagnostics report recovery=none 💬1
  • #81973 [Bug]: Isolated polling ingress never calls bot.init() — Telegram channel halts with “Bot not initialized” 💬3
  • #81963 test(plugin-sdk): track origin-trust SSRF edge coverage and docs 💬1
  • #81962 [Question] Where does documentation feedback go? Testing feedback workflow 💬1

🔒 Closed Issues

  • #77009 feat(bluebubbles): per-DM systemPrompt — follow-up to #60665 (#69198 only addressed groups)
  • #81986 [Feature Request] Message Interruption & Context Merging
  • #81977 [Feature]: Gateway should stop eating client tools
  • #81624 [Bug]: WebSocket multi-account mode — only first account can send group replies (replies=0 for others)
  • #81969 [Bug]: 5.12 strict tool schema injection breaks Xiaomi MiMo custom provider (400 Param Incorrect)
  • #81698 channel-health-monitor doesn’t detect connected-but-frozen Telegram state (sustained getMe fetch-timeouts)
  • #41330 iMessage channel duplicate message loop - messages sent by agent are re-ingested as inbound
  • #41195 Bug: Feishu WebSocket reconnect causes duplicate message processing (dedup cache not restored on restart)
  • #81432 [Bug]: plugin-skills readlinkSync EINVAL on non-symlink directories
  • #81323 [Bug]: channels.openclaw-weixin only enabled:true causes invalid channels.start channel error

hermesagent (NousResearch/hermes-agent)

Stars: 150,479 · Open issues: 11,323 · Last push: <1h ago

✅ Merged PRs

  • #25988 [SANITIZED — possible injection attempt]
  • #25957 fix(ci): stabilize shared test state after 21012
  • #25985 feat: update cron modals
  • #25986 fix(cli): wire /sessions slash command in the classic CLI

🐛 New Issues

  • #26046 kanban: boards rm –delete ineffective when gateway is running (board auto-resurrects)
  • #26045 memory(action=replace) silently clobbers external writes to MEMORY.md (data-loss race with patch tool / shell appends / concurrent sessions) 💬1
  • #26044 [Bug]: Tirith security scanner fails to install on Windows (MSYS/git-bash) - “unsupported_platform”
  • #26042 MCP servers unavailable in cron sessions — ECONNREFUSED when accessing MCP tools from scheduled jobs type/bug comp/cron tool/mcp P2 💬1
  • #26041 QQBot send_message cannot deliver local screenshot media from Windows paths type/bug comp/gateway comp/tools platform/qqbot P3
  • #26037 Feishu: reply-to-image messages lose parent context (_fetch_message_text returns empty for non-text msgs) type/bug comp/gateway platform/feishu P2 💬1
  • #26035 [SANITIZED — possible injection attempt] type/bug comp/agent provider/openai P2
  • #26027 feat: Model routing — use different models for chat vs. tool-use tasks type/feature comp/agent area/config P3
  • #26024 feat(gateway): make busy acknowledgement messages configurable type/feature comp/gateway area/config P3

🔒 Closed Issues

  • #22951 fix(cli): /sessions unknown command in classic REPL mode
  • #21677 Dashboard/TUI slash path reports Unknown command for /sessions

ZeroClaw (zeroclaw-labs/zeroclaw)

Stars: 31,340 · Open issues: 459 · Last push: <1h ago

✅ Merged PRs

  • #6286 fix(channels/telegram): apply mention_only gate to photo, document, and voice updates
  • #6635 fix(cron,channels): carry thread_id through cron announce delivery
  • #6580 fix(provider): honor LM Studio compatible runtime options
  • #6620 fix(tool-call-parser): omit assistant history when no parsed tool calls (#6298)
  • #6631 fix(provider): restrict compatible reasoning effort models
  • #6629 fix(providers): stop replaying stale tool-result images
  • #6598 fix(providers): defensively omit Anthropic temperature for Opus 4.7 (#6147)

🐛 New Issues

  • #6663 feat(telegram): show tool-call progress during partial streaming enhancement risk: medium channel channel:telegram priority:p2 status:accepted status:no-stale
  • #6661 [Feature]: preserve committed streamed output during websocket steering enhancement risk: high agent gateway runtime
  • #6659 No API for pushing notifications into an operator’s gateway session (zeroclaw 0.7.4) bug enhancement risk: high gateway tool gateway: api gateway: ws priority:p1 needs-maintainer-review 💬3
  • #6658 [Feature]: Install script support musl aarch64 linux with –prebuilt bug enhancement risk: medium scripts priority:p2 status:accepted
  • #6657 ci: Advisory scan failed — 2026-05-14 bug type: dependencies risk: high dependencies security priority:p1 status:in-progress status:no-stale 💬1
  • #6654 [Bug]: Cron read-only queries still hit the writable schema-ensure path bug risk: high cron runtime cron:store status:in-progress status:no-stale priority:p3
  • #6653 feat(update): define host-architecture policy for emulated installs enhancement risk: medium core needs-maintainer-review priority:p3
  • #6651 [Bug]: matrix channel leaks ~1 MB Pss per /admin/reload due to upstream Arc cycle in matrix-sdk 0.17 (matrix-rust-sdk#6573) bug risk: medium dependencies channel runtime channel:matrix priority:p2 status:blocked status:no-stale
  • #6648 [Bug]: cron session_target=main still runs in an isolated cron session bug risk: high cron runtime cron:scheduler tool: cron_add priority:p2 status:accepted
  • #6647 [Bug]: Title: Cron job output not routed to configured channels bug risk: high channel cron runtime channel:telegram tool: cron_add priority:p1 status:accepted status:no-stale 💬4
  • #6646 [Bug]: Title: web_search_tool and web_fetch not firing via Telegram channel in v0.7.5 bug risk: high channel provider runtime tool channel:telegram provider:compatible tool: web_search_tool tool: web_fetch priority:p1 needs-maintainer-review status:accepted 💬1
  • #6645 SkillImprover + skill_manage only handle SKILL.toml, not manifest.toml bug risk: high runtime skills tool priority:p2 status:accepted 💬1
  • #6644 Skill review fork: 💾 summary parser misses tool receipts bug risk: low runtime skills tool status:blocked status:no-stale priority:p3 💬1
  • #6642 [Feature]: Capture full prompt/completion on llm.call spans via gen_ai.input.messages / gen_ai.output.messages enhancement risk: medium observability runtime observability: otel observability: runtime_trace priority:p2 status:blocked status:no-stale 💬1
  • #6641 [Feature]: Turn-level OTel trace correlation — nest llm.call / tool.call / memory.* spans under a single turn trace enhancement risk: medium observability runtime observability: otel observability: runtime_trace priority:p2 status:blocked status:no-stale 💬1

🔒 Closed Issues

  • #6229 [Bug]: mention_only=true does not suppress responses to photo/media messages in Telegram groups
  • #6634 [Bug]: cron-scheduled webhook callbacks drop thread_id

IronClaw (nearai/ironclaw)

Stars: 12,250 · Open issues: 934 · Last push: <1h ago

✅ Merged PRs

  • #3556 arch(ws-5): default strategies - Default* impls for all nine traits
  • #3643 arch(ws-3.5): add loop family registry
  • #3553 arch(ws-3): strategy traits gamma - stop, drain, budget
  • #3552 arch(ws-2): strategy traits beta - batch, gate, recovery
  • #3551 arch(ws-1): strategy traits alpha - context, capability, model
  • #3550 arch(ws-0): state, checkpoints, BoundedRing, CapabilityCallSignature, NoProgressDetected
  • #3642 feat(product-workflow): add WebUI inbound DTO contract
  • #3655 fix(reborn): redact loop response output
  • #3656 fix(reborn): flag in-memory durable logs
  • #3657 chore(crates): enable unreachable_pub on tier-2 crates

🐛 New Issues

  • #3673 [QA] openai_compatible provider drops reasoning_content on outgoing requests, breaks DeepSeek v4-pro multi-turn tool calls

🔒 Closed Issues

  • #3624 [Reborn WebUI Beta] Define WebUI inbound DTO contract

🎯 Our Filed Issues (Haderach-Ram on openclaw/openclaw)

  • 🟢 #81567 GPT-4o agent sessions exit after single text response instead of continuing tool-use loop — 💬2 · 15h ago
  • ⚫ #79917 Haderach (OpenClaw bot) permanently auto-banned from OpenClaw Discord — ban reason: “Bot” — 💬3 · 2d ago
  • 🟢 #79607 [Feature]: Identity-based session unification — one session per user regardless of input channel (voice, Telegram, WhatsApp etc.) — 💬1 · 5d ago

🏛️ Official Content — Anthropic + OpenAI

Anthropic — 3 new

OpenAI — 5 new

Next digest: tomorrow 07:45 IST · Radar repo

digest, daily
openclaw ecosystem daily-digest
This post is licensed under CC BY 4.0 by the author.