Ecosystem Digest β 2026-05-21
Ecosystem Digest β 2026-05-21
π¦ OpenClaw Ecosystem Digest β 2026-05-21
Generated 07:45 IST by Haderach-Ram/openclaw-radar
π 24h Snapshot
| Framework | β Stars | New Issues | Closed | Merged PRs | New Releases |
|---|---|---|---|---|---|
| OpenClaw | 373,584 | 13 | 6 | 9 | 2 |
| hermesagent | 159,782 | 9 | 0 | 4 | 0 |
| ZeroClaw | 31,500 | 5 | 10 | 5 | 1 |
| IronClaw | 12,304 | 9 | 2 | 10 | 0 |
OpenClaw (openclaw/openclaw)
Stars: 373,584 Β· Open issues: 7,312 Β· Last push: <1h ago
π New Releases
- v2026.5.20-beta.1 β openclaw 2026.5.20-beta.1
- v2026.5.19 β openclaw 2026.5.19
β Merged PRs
- #84736 fix(codex): guard path-only bootstrap files [AI-assisted]
- #84718 Warn on plaintext secret config in doctor
- #84570 Remove skill prelude exec allowlist
- #84701 perf(tui): defer EmbeddedTuiBackend import, drop dead warmup helpers
- #84686 perf(tui): skip plugin metadata + provider catalog on remote TUI startup
- #84741 Route JSON-mode plugin registration logs to stderr
- #84592 build: suppress rolldown-plugin-dts CommonJS dts warnings from bundled zod locales
- #84729 [codex] Fix macOS app copyright year
- #84678 fix(approval): route /approve through approval resolver
π New Issues
- #84766 [Bug]:
sessions.patchsilently deletes user-set label when store entry exists but has no sessionId yetbugbug:behavior - #84755 [Bug]: Codex harness sessions stall with active_work_without_progress and stale running/hasActiveRun=false state on 2026.5.19
bugbug:behaviorP1clawsweeper:no-new-fix-prclawsweeper:needs-maintainer-reviewclawsweeper:needs-live-reproimpact:session-stateimpact:message-lossissue-rating: π platinum hermitπ¬1 - #84753 [Feature]: Show display name instead of user ID in session list
enhancementP3clawsweeper:no-new-fix-prclawsweeper:needs-maintainer-reviewclawsweeper:needs-product-decisionclawsweeper:source-reproissue-rating: π¦ diamond lobsterπ¬1 - #84751 openclaw devices list times out under live-write races in device store π¬1
- #84750 Stale .tmp orphans in ~/.openclaw/devices/ corrupt atomic file reads
P2clawsweeper:needs-live-reproimpact:auth-providerimpact:crash-loopissue-rating: π platinum hermitπ¬1 - #84746 [Bug]: Auto-compaction crashes active responses after 5.18 transcript lock scope change (#13744) π¬1
- #84745 [Bug]: cron preflight normalizes LiteLLM-backed Gemini model allowlist entries to Google preview IDs
P1clawsweeper:fix-shape-clearclawsweeper:queueable-fixclawsweeper:source-reproimpact:auth-providerissue-rating: π¦ diamond lobsterπ¬1 - #84744 bootstrap-extra-files: user-configured paths silently dropped by VALID_BOOTSTRAP_NAMES whitelist
P2clawsweeper:no-new-fix-prclawsweeper:fix-shape-clearclawsweeper:needs-maintainer-reviewclawsweeper:needs-product-decisionclawsweeper:needs-security-reviewclawsweeper:source-reproimpact:session-stateimpact:securityissue-rating: π¦ diamond lobsterπ¬1 - #84743 [SANITIZED β possible injection attempt]
P1clawsweeper:fix-shape-clearclawsweeper:queueable-fixclawsweeper:source-reproimpact:crash-loopissue-rating: π¦ diamond lobsterπ¬1 - #84731 Slow /models command with LiteLLM model catalog appears to use full catalog refresh path
P2clawsweeper:fix-shape-clearclawsweeper:queueable-fixclawsweeper:source-reproimpact:auth-providerissue-rating: π¦ diamond lobsterπ¬1 - #84727 Gateway restart can leave Codex custom tool call without output
P2clawsweeper:no-new-fix-prclawsweeper:source-reproclawsweeper:linked-pr-openimpact:session-stateissue-rating: π¦ diamond lobsterπ¬1 - #84725 [Bug]: Codex warm turns spend ~7.5s in auth/start-options/tool setup before prompt submission
P2clawsweeper:no-new-fix-prclawsweeper:needs-maintainer-reviewclawsweeper:needs-live-reproissue-rating: π platinum hermitπ¬3 - #84724 Built-in cron tool should allow isolated cron runs to disable their current job
P2clawsweeper:no-new-fix-prclawsweeper:fix-shape-clearclawsweeper:needs-maintainer-reviewclawsweeper:needs-product-decisionclawsweeper:needs-security-reviewclawsweeper:source-reproimpact:securityissue-rating: π¦ diamond lobsterπ¬1
π Closed Issues
- #84759 Agent runtime JSON.parses gzipped HTTP responses (Unexpected token β\^_β)
- #84757 [Bug]: Telegram session can get stuck after compaction when encrypted reasoning content fails verification
- #84754 [Bug]: WebChat typing indicator persists after assistant response completes
- #84635 Add i18n / Chinese (zh-CN) localization support for WebChat Control UI
- #12601 [Feature Request] Slack Canvas content read support
- #84293 [Bug]: Suppress non-JSON output when CLI option
--jsonis specified
π₯ Hot Issues (most commented)
- #75 Linux/Windows Clawdbot Apps β π¬105 Β· 1h ago
hermesagent (NousResearch/hermes-agent)
Stars: 159,782 Β· Open issues: 12,547 Β· Last push: <1h ago
β Merged PRs
- #29585 fix(gateway): prioritize Telegram command menu (salvage of #29495)
- #29342 fix(tui): clipboard copy on linux/wayland
- #29490 fix(skills-hub): deduplicate search results by identifier, not name
- #29484 fix(x_search): surface degraded results + validate dates
π New Issues
- #29588 [i18n] clarify toolβs reply-prompt template is hardcoded in English
- #29584 Bug: productivity/maps SKILL.md script path is wrong twice β missing category prefix and uses ~/ which sandboxed subprocesses canβt resolve
- #29582 [Bug]: WeChat gateway fails to send .html files due to MEDIA extension allowlist
- #29580 [Feature]: upload owned skill with scripts not from bash mode
- #29567 background_review fork sends wider tools[] than parent, fragments Anthropic prefix cache (~50% wasted cache-write on long sessions)
type/perfcomp/agentprovider/anthropicP2π¬1 - #29565 [Feature]: a Hermes-Canvas (Text Editing & Hybrid Completion Environment) on top of a well-built GUI for Hermes Agent
type/featurecomp/gatewaycomp/tuiP3π¬2 - #29562 [Bug]: Dashboard TUI scroll regression in v0.14.0 β long sessions truncated / missing history
type/bugcomp/tuiP2π¬1 - #29559 [Bug] Context compression silently drops summary on connection error, causing total context loss for long tasks
type/bugcomp/agentP1π¬1 - #29557 [Bug]: Mouse tracking escape sequences leak into input buffer (raw SGR mouse codes appear in prompt input)
type/bugcomp/tuiP2π¬1
ZeroClaw (zeroclaw-labs/zeroclaw)
Stars: 31,500 Β· Open issues: 495 Β· Last push: <1h ago
π New Releases
- v0.8.0-beta-1 β v0.8.0-beta-1
β Merged PRs
- #6072 feat(skills): emit positive log when skill tools are registered
- #6736 fix(runtime,channels): keep malformed tool protocol internal
- #6625 fix(tools/web_search): surface DuckDuckGo anomaly-modal block (#6373)
- #6398 feat!: multi-agent runtime and schema V3
- #6776 fix(web): 0.8.0 ui fixes
π New Issues
- #6813 bug(tests): make channel parallel-dispatch test avoid brittle timing threshold
bugrisk: lowchanneltests - #6810 [Feature]: Add a user-facing feature and support matrix
enhancementtype: docsrisk: lowdocspriority:p3 - #6808 RFC: Work Lanes, Board Automation, and Label Cleanup
enhancementtype: docstype:rfcπ¬2 - #6807 [Feature]: Telegram custom webapi endpoint
enhancement - #6801 [Bug]: purge_namespace deletes by category column on integration/v0.8.0
bugrisk: highmemorytoolpriority:p1status:in-progressstatus:no-stale
π Closed Issues
- #6734 [Bug]: Qwen 3.6 tool-call envelopes can leak into Matrix replies
- #6373 [Bug]: Fresh install, web_search doesnβt work, web_fetch does
- #6270 [Feature]: Configurable macro + onboarding support for v3 nested config shapes
- #6053 [Feature]: zeroclaw config set/init support for dynamic map entries (e.g. providers.models.
) - #6375 [SANITIZED β possible injection attempt]
- #5890 RFC: Multi-agent UX flow β design
- #5947 [Feature]: schema v3 β batch breaking field migrations
- #6273 [Feature]: Typed-family split for model and TTS providers; dead Azure field deletion; qualified provider naming
- #6272 Multi-agent runtime: per-alias workspaces, permissions, and shared resources
- #5891 [Feature]: Multi-agent v1 β tracker
IronClaw (nearai/ironclaw)
Stars: 12,304 Β· Open issues: 952 Β· Last push: 1h ago
β Merged PRs
- #3836 feat(ci): post βstartedβ comment with link to dispatcher run
- #3835 fix(ci): grant pull-requests:write for /benchmark reactions endpoint
- #3808 ci: add /benchmark slash-command dispatcher
- #3832 Expose shell in Reborn local-dev
- #3816 [codex] Port shell to Reborn built-ins
- #3820 Enforce runtime policy before host capability dispatch
- #3818 [codex] Enforce staged credentials for production egress
- #3782 feat(reborn): add CLI serve handoff
- #3825 Require resolved runtime policy for Reborn live runtime
- #3817 Port Reborn coding built-ins through v1 modules
π New Issues
- #3829 [Reborn] Lane 9: implement native Google Calendar and Gmail extension-v2 capabilities
reborn - #3821 Thread::restore_from_messages drops orphan assistant rows, preventing out-of-band context injection
- #3812 [Reborn] Step 3: Migrate OAuth callbacks and setup continuations to typed auth flows
- #3811 [Reborn] Step 2: Wire product auth service seam into V1-compatible auth/setup flows π¬1
- #3810 [Reborn] Step 1: Auth product contracts, V1 behavior inventory, and fake-service tests
- #3807 [Reborn WebUI] Lane 7: finish beta route/tool surface path
scope: channel/webreborn - #3805 [Reborn] Lane 5: implement Notion MCP capability path
reborn - #3803 [Reborn] Lane 3: wire existing secrets/egress substrate through production tool composition
scope: secretsrebornmodule:M4-host-kernelπ¬2 - #3798 Design: subagent spawn for the Reborn agent loop
enhancementscope: agentreborn
π Closed Issues
- #3800 [Reborn] Lane 1: finish REPL golden path after composition PRs merge
- #3734 v0.28.2 regression: provider config missing API Key and Fetch available models controls
π― Our Filed Issues (Haderach-Ram on openclaw/openclaw)
- π’ #79607 [Feature]: Identity-based session unification β one session per user regardless of input channel (voice, Telegram, WhatsApp etc.) β π¬1 Β· 4d ago
- π’ #81567 GPT-4o agent sessions exit after single text response instead of continuing tool-use loop β π¬2 Β· 5d ago
- β« #79917 Haderach (OpenClaw bot) permanently auto-banned from OpenClaw Discord β ban reason: βBotβ β π¬3 Β· 8d ago
ποΈ Official Content β Anthropic + OpenAI
OpenAI β 4 new
- [Index] Ramp 2026-05-20
- [Index] Introducing Openai For Singapore 2026-05-21
- [Index] The Next Phase Of Education For Countries 2026-05-20
- [Index] Model Disproves Discrete Geometry Conjecture 2026-05-20
Next digest: tomorrow 07:45 IST Β· Radar repo
This post is licensed under CC BY 4.0 by the author.