🚀 Astron Open Source · Bi-Weekly Report (Aug 24 – Sep 6, 2026)

[FenjuFu]

🚀 Astron Open Source · Bi-Weekly Report (Aug 24 – Sep 6, 2026)

---

🌟 Highlights

• The Spark-X2.5 open model series launched globally on Sep 1 — Spark-X2.5-4B and Spark-X2.5-1.7B, Apache-2.0, with a hybrid-attention architecture, a native 1M-token context, and support for 200+ languages. Day-0 serving support arrived from SGLang and vLLM, and the models shipped simultaneously to Hugging Face, ModelScope, Ollama, Modelers and SCNet.
• Ten Spark-X2.5 model repositories and one dataset are now public on Hugging Face, together accounting for roughly 74,000 downloads and 1,096 likes at the time of writing. Spark-X2.5-4B reached the Hugging Face 7-day trending models board during the cycle.
• Thirteen repositories across the ecosystem merged 138 PRs: SkillHub 62 · Astron Agent 30 · Dolphin MCP Pilot 12 · MemFlywheel 6 · Domux 5 · Astron RPA 2 · AstronClaw Tutorial 1 · HarnessClaw 1 · and 19 across the XHToken repositories.
• Six releases landed: SkillHub v0.2.18 and v0.2.19 plus CLI 0.1.10 and 0.1.11, Dolphin MCP Pilot v0.3.1, and the first vllm-spark2_5-plugin v0.1.0.
• Astron Agent spent the cycle on security hardening. Thirty merged PRs covering the code-node executor, SSRF, XSS/artifact/sandbox boundaries and internal service authentication were released as the v1.1.2 security release on Sep 7, one day after this reporting window closed. Operators on v1.1.1 or earlier should upgrade.
• Three community-reported SkillHub bugs were filed, fixed and closed inside the same two weeks — the disabled download button on promoted global Skills, the missing pagination on the promotion review page, and a stale SESSION cookie that blocked re-login.
• Two challenges ran in parallel: HER Hack-Astron 讨论：awesome 站 Demo 导航中 astron-agent 指向 Zeabur 部署入口，是否继续长期维护此类开源 Demo？ #4 (Domux) completed judging on Sep 2, and HER Hack-Astron 🚀 Astron Open Source · Bi-Weekly Report (Jun 15 – Jun 28, 2026) #5 (Spark-X2.5) closed submissions on Sep 6 with judging scheduled for Sep 7–13.

👥 Top 10 Active Community Contributors

Based on message volume across 25 exported open-source discussion-group records. Automated Q&A and welcome accounts were excluded; display-name aliases were normalized and nicknames anonymized. Thank you to everyone who answered questions, asked questions, and built alongside us 🙌

---

📦 Release & Development Highlights

✨ Spark-X2.5 — a compact on-device model series, open-sourced

Spark-X2.5-4B and Spark-X2.5-1.7B are general-purpose models built for practical, efficient deployment rather than benchmark-only appeal.

• 🧠 Efficient architecture with a native 1M-token context. The models combine one full-attention layer with three sliding-window attention layers, which cuts the overhead usually associated with long context while still supporting up to 1M tokens natively.
• 🤖 Coding and agent capability at a small size. The models integrate with common agent harnesses including Codex, Claude Code, OpenClaw and Hermes.
• 🖥️ Broad hardware and framework coverage. NVIDIA, Huawei Ascend, Hygon and HOUMO.AI hardware; vLLM, SGLang, llama.cpp and MLX inference; Ollama and LM Studio for quick local deployment; LLaMA-Factory for fine-tuning.
• 📅 Release timeline inside the window: global launch across five distribution platforms (Sep 1) → AtomGit added as a channel (Sep 2) → vLLM and SGLang deployment support on Ascend NPU (Sep 3) → FP8 and INT8 quantized weights for both sizes (Sep 4) → native Spark2_5ForCausalLM support merged upstream into llama.cpp b10829 (Sep 6).

Supporting repositories moved with the model. vllm-spark2_5-plugin v0.1.0 (Aug 28) brings Spark2_5 to vLLM as an out-of-tree plugin with no upstream patches, registering a spark25 tool-call parser that emits typed JSON, and shipping with Apache-2.0 licensing, third-party attribution, a lint/test/build CI pipeline and automated PyPI publishing. Spark-MLX-LLM added Spark2_5 inference for MLX plus tool wrappers, the XHToken llama.cpp fork landed the tokenizer hash, auto parser and chat-template work, and LlamaFactory gained a fine-tuning template and guide.

🧰 SkillHub v0.2.18 and v0.2.19 — personal namespaces, workspace sync, and a reviewable publishing loop

SkillHub merged 62 PRs, the highest of any repository this cycle, and published two server releases plus two CLI releases.

v0.2.18 (Aug 29)

• 🏷️ Personal namespaces can be auto-provisioned for newly activated accounts, with operator-configurable naming and idempotent ownership handling (#712).
• 🔄 CLI workspace synchronization arrived in CLI 0.1.10: sync pull, status, diff and push, backed by incremental fingerprint metadata (#724). Follow-ups made fingerprints deterministic across server and Node path ordering, so a clean mixed-path workspace is no longer reported as locally changed (#780).
• 📂 Folder publishing in the Web UI: the browser builds a validated ZIP client-side, filtering VCS, build and OS metadata, with no new dependency (#757).
• 🧱 linux/riscv64 Server and Web images, with the Scanner and the full Compose dependency stack explicitly out of scope (#725); Russian localization was contributed by @michael-xiii (#700).
• 🛡️ Transactional scan-task outbox: scan tasks can no longer be lost between the database commit and Redis publication, with retries, reclaiming, cleanup and consumer idempotency (#752).
• 🐳 Upgrade safety: the Server's numeric UID/GID is preserved across the Alpine-to-Noble runtime-image change, so local-storage volumes created by v0.2.17 stay writable (#779). Flyway V44/V45 apply automatically — back up PostgreSQL first.

v0.2.19 (Sep 4)

• 📝 The review loop became visible to authors: review progress, durable attempt history, and resubmission of rejected versions (#791), plus public text reviews with author editing and audited administrator moderation (#793).
• 🔁 Failed security scans became recoverable: a permission-checked, idempotent retry for terminal SCAN_FAILED versions (#811), and bounded recovery so expired deliveries become SCAN_FAILED rather than pending forever (#809).
• 🧭 Install-for-agent prompt on skill details, preserving full namespace coordinates (#788), and the pinned HOL Guard plugin-scanner Skill added to the curated starter collection — a Skill, not an extension of the Server scanner pipeline (#814).
• ⚙️ CLI 0.1.11 adds bounded source-safe Skill upgrades, fingerprint-safe install and sync, explicit namespace pull selection, and accurate asynchronous submission status (#796, #804, #815, #816).
• 🐛 Correctness work included deterministic conflicts for concurrent publishes of the same coordinate (#784), converged concurrent OAuth identity binding and atomic governance-plus-audit mutations (#786, #787), paginated promotion review queues (#774), and SPA entry revalidation after deployments (#812).

🤖 Astron Agent — a security-hardening cycle, released as v1.1.2

Astron Agent merged 30 PRs during the window. Almost all of them fed into the v1.1.2 security release published on Sep 7, one day after this window closed.

• 🚨 Workflow code execution and tenant isolation. Versions <= v1.1.1 using the legacy unsandboxed local code executor allowed an authenticated user with workflow-creation permission to execute code inside the workflow service container and, in affected deployments, reach other tenants' data. v1.1.2 rejects unsafe local execution and defaults to the built-in isolated LangChain/Pyodide executor, migrating untouched legacy configuration while preserving explicit process-level overrides (#1650, #1651). Advisory: GHSA-mh3w-4q3f-2fg5. Reported by Noy Pearl and the Lasso Security team.
• 🔒 SSRF prevention in outbound plugin requests (#1669) and hardened XSS, artifact and sandbox boundaries with scoped CSRF exclusions (#1643).
• 🐳 A secure deployment contract enforced in the Docker layer (#1633), MinIO policy compatibility preserved (#1634), and static migration statements in the console (#1640).
• 🧩 Workflow runtime fixes: isolated code fallback enabled by default (#1649) and zero-parameter code entrypoints supported (#1636).
• 📚 Documentation and build reliability: a FAQ entry for workflow publishing failing when basic info is missing (#1630), a Related Projects section linking sibling Astron repositories (#1627), a repaired Japanese README image path (#1653), a VitePress example gallery fix (#1655), and PyPI/Maven Central retry handling (#1660, #1657).

⚠️ Upgrade note. CODE_EXEC_TYPE=local is no longer supported. Update deployment files and service images together — updating only the source checkout or only the workflow image does not apply the complete fix.

⚙️ Astron RPA — log sanitization and a repaired MCP discovery chain

Astron RPA merged 2 PRs, both closing reported defects.

• 🔐 Sensitive data could reach OpenAPI, rpa-auth and nginx logs (#858, closing #855). A unified sanitizer now redacts API keys and secrets, bcrypt hashes, Authorization/Token/Cookie/Session values, passwords, verification codes, temporary credentials and phone numbers across application logs, Uvicorn access/error logs, Logback output and stack traces. The nginx access-log format switched to $uri without query parameters and dropped Referer, while keeping method, status, size, User-Agent and latency for troubleshooting.
• 🔗 Published workflows were invisible to MCP tools/list (#860, closing #859). robot-service synced workflows to a container address that does not exist in the official Compose stack, and omitted active-status and name metadata. The sync URL is now configurable via OPENAPI_WORKFLOWS_UPSERT_URL with the old address retained as a fallback, and the published-workflow DTO carries project_id, name, description, version, status = 1 and parameters.

🐬 Dolphin MCP Pilot v0.3.1 — discoverable through the official MCP Registry

Dolphin MCP Pilot merged 12 PRs and published v0.3.1 on Aug 31.

• 📇 Published to the official MCP Registry as io.github.iflytek/dolphin-mcp-pilot, so MCP hosts can find and install it without being pointed at the repository by hand (#20). Publication is gated on the image push, authenticates with GitHub OIDC, validates server.json against its schema, and re-checks the GHCR ownership label before publishing.
• 🗂️ A community case-study gallery under cases/ collected six accepted studies covering release-gate recovery, safe weekday schedule changes, human-gated project cleanup, request-scoped team isolation, and nightly-ETL incident firefighting (#24, #39, #40, #41, #46).
• 📦 PyPI trusted publishing was added (#44); until it lands, Docker and source remain the supported install paths.

🏠 Domux — the HER Hack-Astron #4 case corpus

Domux merged 5 PRs: the case framework with automatic validation (#15) plus four community case studies (#23, #24, #26, #29). See the Community Spotlight below.

---

💬 Community Q&A & Discussions (Bi-Weekly Picks)

The following themes were distilled from anonymized WeCom support and open-source discussion records for Aug 24 – Sep 6. Automated Q&A answers were not treated as authoritative unless a human resolution or repository evidence confirmed them — several were wrong in this cycle and were corrected in-thread by maintainers or by other community members. 🟢 Resolved · ℹ️ Clarified · 🟡 Follow-up needed.

Spark-X2.5

• ℹ️ What hardware does it actually need? For a single stream at the full 1M context, plan on about 32 GB of VRAM; 128 K context runs comfortably on 16 GB. CPU-only is viable for the 1.7B: community members measured roughly 2–3 tokens/s for BF16 weights on a 16 GB business laptop, and about 24 tokens/s prefill with 9.5 tokens/s decode for a quantized build on 8 CPU cores. Non-NVIDIA accelerators including Ascend and Hygon are supported.
• 🟢 Ollama and LM Studio need a local build for now. Until upstream support landed, running Spark-X2.5 through Ollama or LM Studio required compiling the XHToken llama.cpp fork and pointing the runtime at it — a macOS LM Studio load error in one thread was resolved exactly this way. Community members shared the working build path, and a contributor noted you can skip the manual GGUF download entirely with ollama run SparkLLM/Spark-X2.5-4B. Native Spark2_5ForCausalLM support then merged upstream into llama.cpp b10829 on Sep 6.

• ℹ️ Does it conflict with an existing Ollama install? Shut the stock Ollama service down first; the Spark-enabled build then serves on the same 11434 port, so other applications keep working against the familiar endpoint.
• ℹ️ Is there a GGUF build of the 4B? Yes — XHToken/Spark-X2.5-4B-GGUF. The Q&A bot answered that only the 1.7B had GGUF weights; a community member supplied the correct link in the thread. FP8 and INT8 weights followed on Sep 4, with GGUF builds of those quantizations expected the following week.
• ℹ️ How much does 4-bit quantization cost you? One contributor benchmarked IQ4_NL, IQ4_XS, XS-UD and Q4_K_M against the BF16 baseline on their own instruction-following task set and measured 94.5%–95.8% retention, using an imatrix calibrated for that specific task — a task-tuned calibration set, not a general-purpose claim.

• 🟡 The 1.7B can fall into repetition loops on some prompts where the 4B stays stable; this was independently observed by more than one person and by the team. Separately, the SGLang nightly image referenced in the documentation did not have the hybrid sliding-window attention pool enabled — the fix was committed to the repository but had not yet reached a published image at the time of the thread.
• ℹ️ Is there a vision or multimodal variant? No. Spark-X2.5 is text-only.
• ℹ️ Downloading from mainland China: use the ModelScope mirror; AtomGit, Modelers and SCNet are also official channels.

SkillHub

• 🟢 Why is the download button greyed out on a Skill promoted to @global? This was reported on Aug 27 by a member who noticed the behaviour contradicted the documented rule that @global + PUBLIC allows anonymous search, inspect and download. The maintainer confirmed it as a bug rather than intended design, asked for an issue, and #763 was closed on Sep 2 with regression coverage added in #792.

• 🟢 Older pending promotion requests disappeared from the review page. The administrator promotion queue rendered without pagination, so once enough requests accumulated the earlier ones became unreachable. Reported Aug 28 as #772, closed Aug 31, with paginated review queues shipping in v0.2.19 (#774).
• 🟢 A stale SESSION cookie blocked re-login with "Session expired" until cookies were cleared (#802, closed Sep 3). The fix keeps public routes accessible when a browser carries an expired session (#783) alongside broader session-boundary hardening (#801); a member asked whether #801 was in the latest release — it shipped in v0.2.19 on Sep 4.
• ℹ️ When is LDAP support landing? A contributor pointed out that the roadmap item was still unchecked even though #672 looked merged. The maintainer clarified that it merged into a feature branch rather than main, that an earlier community PR was not general enough, and that the unified identity design needs another pass before it ships.
• ℹ️ Will SkillHub keep tracking the ClawHub CLI? No. Chasing every ClawHub release is not sustainable, so the recommendation is the SkillHub CLI from npm. v0.2.19 documents which ClawHub search, inspect, install and publish workflows remain compatible (#806).
• 🟡 skill-scanner reports "scan failed" with 0 findings. v0.2.19 makes terminal SCAN_FAILED versions retryable (#811) and bounds recovery of unavailable tasks (#809), so the state is no longer a dead end. The underlying cause is still per-deployment: capture the scanner worker logs and the version's audit record before escalating.

Astron RPA

• 🟢 An MCP call returns code: O000 / "运行成功" but data: {}. Fixed by #860 on Sep 3. The published workflow was never fully registered with the OpenAPI service, so MCP could report success without the workflow being discoverable or returning output. Upgrade to a build containing the fix, and set OPENAPI_WORKFLOWS_UPSERT_URL if your deployment does not use the official Compose service names.
• ℹ️ The EXE installs but the client shows 502 on launch. The desktop client is not a standalone application — it requires reachable backend services. A 502 means the client started but could not reach a working server. Note also that the GitHub release artifact and the client published on the official site are not interchangeable.
• ℹ️ Is there a web console in the open-source edition? No. The open-source stack is desktop-client based and currently targets Windows 10/11.
• ℹ️ Does it support scheduled execution? Yes — direct run, scheduled tasks with cron expressions, scheduler mode and API invocation are all available.
• ℹ️ Fingerprint-browser integration for multi-account scenarios? Not currently supported. The browser automation path is built on the browser plugin architecture for Chrome, Edge and Firefox; there is no documented way to attach to a fingerprint browser or bind through the Chrome DevTools Protocol.
• ℹ️ Is there a KingbaseES (人大金仓) Docker image? Not provided. MySQL and Redis are the documented defaults for server deployment.
• 🟡 Nightly runs fail with "user actively closed" although nobody touched the machine. The record did not establish a root cause. Before escalating, capture the client and scheduler logs around the failure timestamp, the session/idle timeout configuration, and whether the host sleeps, locks or drops its network connection overnight — treat "user actively closed" as a symptom label, not a diagnosis.

Astron Agent

• ℹ️ Publishing a workflow fails without a clear reason. The most common cause is incomplete basic information on the workflow; the cycle added an FAQ entry covering it (#1630).
• 🚨 If you self-host, upgrade. See the v1.1.2 notes above. Automated vulnerability scanners may not flag this application-level issue, so verify deployed image versions explicitly rather than relying on a scan result.

---

🔭 Community Spotlight: HER Hack-Astron #4 and #5

HER Hack-Astron #4 — Domux smart-home real-world cases (judged Sep 2)

Contributors ran real inference, evaluation, quantization, training or integration experiments against Domux, published the full write-up in the model's Hugging Face Discussions, and submitted it as a PR into cases/.

🏆 Champion — domux-seniorsafe (Discussion #7, PR #29 by @qrx-joe). It won not for the best-looking numbers. It was the only entry that completed a three-round freeze-then-run-then-publish progression: after fixing a normalization defect it re-ran the entire dataset rather than cherry-picking, lifting normalized exact match from 75.7% to 85.7%, and proved both runs used the same model and configuration by showing 80/80 raw-pipeline outputs were byte-identical to round one. Round three froze a 160-case challenge set, the evaluation protocol and a SHA-256 fingerprint before running 320 real CPU inferences — then published the unflattering results in full (raw 55.1%, normalized 61.8%, 5/24 input-strategy misses, 2/24 blocked-label requests still surfacing as candidates) and explicitly declined to patch defects discovered after the freeze. The case also labelled its own "100% rule-layer safety" figure as true by construction, not a model metric, dismantling a number that could easily have been presented as capability.

🌟 Outstanding contributions: domux-4057-full-eval (full 4,057-sample BF16 vs NF4 comparison on an RTX 4090D, publishing 97.58% / 93.57% accuracy and the "saves memory but does not speed things up" deployment boundary); domux-4057-rtx3090-eval (PR #23, consumer-GPU reproduction reporting accuracy, concurrent latency and single-request latency separately); domux-safety-gate (PR #24, advancing from input-only checks to all seven output fields and keeping 18/28 false interventions in the report); domux-clarify-commit (PR #26, 96 BF16 inferences driving real Home Assistant REST state changes, 3/3 exact commits and 1/1 pre-dispatch rejection of genuine state drift); and domux-mcp-server (a four-tier auth, high-risk confirmation, audit-log and risk-scoring integration direction).

Two award-winning entries were deliberately not merged: an award does not waive the repository's hard acceptance criteria — one failed the account-reputation check, the other had an unsigned DCO and was missing the source, tests and raw evaluation artifacts its own reproduction commands required.

HER Hack-Astron #5 — Spark-X2.5 edge-agent cases (submissions closed Sep 6)

The fifth challenge opened the same day the model launched. Every entrant from #4 was personally invited to carry their frozen-evaluation, CPU/GPU deployment, failure-analysis and safety-gate methods into the new subject matter.

Submissions landed across a wide hardware and topic range: tool-calling safety boundaries on an Apple M3; Q4_K_M coding, throughput and OOM boundaries on a discrete-GPU-free 8-core/16 GB Linux box; Chinese meeting-note JSON extraction on a 24 GB laptop CPU; multilingual (English/Chinese/Thai) guardrails on an Intel Mac showing that valid JSON can still be unsafe; a bilingual CPU safety gate for destructive automation; a Godot game prototype driven from bad JSON to a playable hundred-round build; an OpenClaw agent tool-calling and safety evaluation scoring 20/21; agent-handoff continuity testing on an M3; edge deployment on an 8 GB AMD RX 580 over Vulkan; and a repository-review case that traced a vLLM 0.23 parser compatibility break through three fixed-protocol reviews and landed the fix upstream as Spark-plugin #10.

Submissions closed at 24:00 Beijing time on Sep 6 and judging runs Sep 7–13, so this report does not announce a winner.

👉 HER Hack-Astron #5 · Spark-X2.5 edge-agent challenge

Hack-Astron x openKylin #1 — running skills on a domestic OS (Sep 3 – 24)

A new challenge opened on Sep 3, jointly with openKylin. Entrants deploy the open-source editions of Astron Agent and SkillHub locally on openKylin Desktop 2.0+, implement skill logic with Astron Agent's native RPA engine, and submit through a PR to the openKylin-skills repository. Cloud-only submissions without a completed local deployment are not accepted. Suggested directions include UKUI desktop automation, package management, community-operations automation, WPS/LibreOffice RPA and system health inspection. Contest runs to Sep 24, judging Sep 25–30.

👉 Hack-Astron x openKylin #1

Community momentum

The @SparkLLM account passed 500 followers within 12 hours of the launch, and independent reviews, videos and mobile-integration experiments — including PocketPal AI with the XHToken llama.cpp fork for iOS and Android local inference — appeared within days.

---

🙌 Build With Us

Good starting points from this cycle include:

• run Spark-X2.5 on hardware we have not covered yet — Ascend, Hygon, older consumer GPUs, or plain CPU — and publish throughput, memory and failure boundaries rather than a single score;
• contribute a reproducible quantization comparison with a general-purpose calibration set, so the community has a counterpart to the task-tuned measurements published so far;
• reproduce the 1.7B repetition-loop behaviour with a minimal prompt and a fixed revision, and file it with raw outputs;
• validate SkillHub's namespace workspace sync and the new review/resubmission loop against a realistic multi-reviewer setup;
• test the Astron Agent v1.1.2 isolated code executor against workflows that previously relied on host filesystem, subprocess or network access, and report what needs adapting;
• contribute a redacted Dolphin MCP Pilot case that changes real workflow state and verifies the terminal outcome;
• build an openKylin skill end-to-end on a local Astron Agent + SkillHub deployment for Hack-Astron x openKylin Welcome to the iFLYTEK Open Source Community 👋 #1.

Before starting, leave a comment on the relevant issue so maintainers can confirm scope and avoid duplicated work.

---

🔗 Explore the Astron Open-Source Ecosystem

• Astron Agent — enterprise-grade agentic workflow platform
• SkillHub — self-hosted registry and governance platform for Agent Skills
• Astron RPA — Agent-ready RPA for desktop and web automation
• Dolphin MCP Pilot — production MCP server for Apache DolphinScheduler
• Domux — low-latency smart-home command parsing model
• MemFlywheel — agent-native long-term memory
• iFly-Skills · AstronClaw Tutorial · HarnessClaw
• Spark-X2.5 and the XHToken organization · Spark-X2.5 on Hugging Face
• All repositories tagged iflytek-astron

Methodology note: PR totals use GitHub merged_at timestamps from Aug 24 through Sep 6, 2026, across the eleven repositories tagged iflytek-astron plus the XHToken organization repositories. Release dates use GitHub published_at. Hugging Face download and like counts are a snapshot taken on Sep 8, 2026 and will keep moving. Astron Agent v1.1.2 was published on Sep 7, one day after the window closed, and is reported here because the work it contains was merged inside the window. Community questions were anonymized and condensed; screenshots reproduced here were shared in the open-source discussion groups, and images containing identifiable account, host or organization details were excluded. Customer and group identities and environment-specific details were not published.

Questions, ideas, or a contribution to share? Join the community discussion or open an Issue / PR in the relevant repository. Let's keep building open, production-ready infrastructure for enterprise AI Agents 🚀

---

🚀 Astron 开源双周报（2026 年 8 月 24 日 – 9 月 6 日）

---

🌟 亮点速览

• 星火 Spark-X2.5 开源模型系列于 9 月 1 日全球发布 —— Spark-X2.5-4B 与 Spark-X2.5-1.7B，Apache-2.0 协议，采用混合注意力架构，原生 100 万 Token 上下文，支持 200+ 语言。SGLang 与 vLLM 均提供 Day-0 推理支持，模型同步上线 Hugging Face、魔搭 ModelScope、Ollama、Modelers 与 SCNet。
• Hugging Face 上现已公开 10 个 Spark-X2.5 模型仓库与 1 个数据集，合计约 7.4 万次下载、1,096 个 Like；本周期内 Spark-X2.5-4B 登上 Hugging Face 近 7 天模型趋势榜。
• 生态内 13 个仓库共合并 138 个 PR：SkillHub 62 · Astron Agent 30 · Dolphin MCP Pilot 12 · MemFlywheel 6 · Domux 5 · Astron RPA 2 · AstronClaw Tutorial 1 · HarnessClaw 1，以及 XHToken 各仓库合计 19 个。
• 本周期发布 6 个版本：SkillHub v0.2.18、v0.2.19 及 CLI 0.1.10、0.1.11，Dolphin MCP Pilot v0.3.1，以及首个 vllm-spark2_5-plugin v0.1.0。
• Astron Agent 本周期集中做安全加固：30 个合并 PR 覆盖代码节点执行器、SSRF、XSS/制品/沙箱边界与内部服务鉴权，最终以 v1.1.2 安全版本于 9 月 7 日发布（统计窗口结束后一天）。运行 v1.1.1 及更早版本的部署方应尽快升级。
• 3 个社区反馈的 SkillHub 缺陷在同一个双周内被提出、修复并关闭 —— 全局推广后 Skill 下载按钮置灰、推广审核页缺少分页、过期 SESSION Cookie 导致无法重新登录。
• 两场挑战赛并行推进：HER Hack-Astron 讨论：awesome 站 Demo 导航中 astron-agent 指向 Zeabur 部署入口，是否继续长期维护此类开源 Demo？ #4（Domux）已于 9 月 2 日完成评选；HER Hack-Astron 🚀 Astron Open Source · Bi-Weekly Report (Jun 15 – Jun 28, 2026) #5（Spark-X2.5）于 9 月 6 日截止提交，评审窗口为 9 月 7—13 日。

👥 社区最活跃 Top 10 贡献者

基于 25 个开源讨论群导出记录的消息量统计（已排除问答 bot、欢迎 bot 等自动账号，并归一同一人的展示名）；昵称已匿名处理。感谢每一位答疑、提问并与我们共同建设的朋友 🙌

---

📦 发布与开发进展

✨ Spark-X2.5：端侧小参数模型系列开源

Spark-X2.5-4B 与 Spark-X2.5-1.7B 是面向真实部署而非仅面向榜单的通用模型。

• 🧠 高效架构 + 原生 1M 上下文：1 层全注意力 + 3 层滑动窗口注意力的混合结构，在原生支持 100 万 Token 上下文的同时显著降低长上下文常见的计算开销。
• 🤖 小体量下的编码与 Agent 能力：与 Codex、Claude Code、OpenClaw、Hermes 等主流 Agent harness 深度适配。
• 🖥️ 广泛的软硬件兼容：支持 NVIDIA、华为昇腾、海光、后摩智能等硬件；支持 vLLM、SGLang、llama.cpp、MLX 推理；可通过 Ollama、LM Studio 快速本地部署；可用 LLaMA-Factory 微调。
• 📅 窗口内发布节奏：五大平台全球发布（9/1）→ 新增 AtomGit 分发渠道（9/2）→ vLLM 与 SGLang 昇腾 NPU 部署支持（9/3）→ 发布 FP8 与 INT8 量化权重（9/4）→ Spark2_5ForCausalLM 原生架构支持合入 llama.cpp 上游 b10829（9/6）。

配套仓库同步推进。vllm-spark2_5-plugin v0.1.0（8/28）以 out-of-tree 插件形式为 vLLM 带来 Spark2_5 支持，无需修改上游源码，注册名为 spark25 的工具调用解析器并输出强类型 JSON；同时具备 Apache-2.0 许可与第三方声明、lint/test/build CI 流水线和 PyPI 自动发布。Spark-MLX-LLM 新增 Spark2_5 的 MLX 推理与工具封装，XHToken 的 llama.cpp fork 完成 tokenizer hash、auto parser 与 chat template 适配，LlamaFactory 补充了微调模板与说明。

🧰 SkillHub v0.2.18 与 v0.2.19：个人命名空间、工作区同步与可见的审核闭环

SkillHub 本周期合并 62 个 PR（全生态最高），并发布两个服务端版本与两个 CLI 版本。

v0.2.18（8 月 29 日）

• 🏷️ 个人命名空间自动创建：新激活账号可自动开通个人 namespace，命名规则由运营方配置，归属处理具备幂等性（#712）。
• 🔄 CLI 工作区同步：CLI 0.1.10 新增 sync pull、status、diff、push，并引入增量指纹元数据（#724）。后续修复令指纹在服务端与 Node 路径排序下保持确定，避免干净的混合路径工作区被误报为本地已修改（#780）。
• 📂 Web 端直接选择文件夹发布：浏览器侧构建经校验的 ZIP，自动过滤 VCS、构建与操作系统元数据，且未引入新依赖（#757）。
• 🧱 Server / Web 镜像新增 linux/riscv64，并明确 Scanner 与完整 Compose 依赖栈不在支持范围内（#725）；俄语本地化由 @michael-xiii 贡献（#700）。
• 🛡️ 扫描任务事务型 Outbox：扫描任务不再在数据库提交与 Redis 发布之间丢失，并补充重试、重认领、清理与消费幂等（#752）。
• 🐳 升级安全性：Alpine 迁移到 Noble 运行时镜像后仍保留 Server 的数值 UID/GID，v0.2.17 创建的本地存储卷升级后依然可写（#779）。Flyway V44/V45 会自动执行，升级前请先备份 PostgreSQL。

v0.2.19（9 月 4 日）

• 📝 审核过程对作者可见：新增审核进度、可持久化的提交尝试历史，以及被驳回版本的重新提交（#791）；同时支持公开文字评价、作者自行编辑与带审计的管理员管控（#793）。
• 🔁 扫描失败可恢复：终态 SCAN_FAILED 版本支持带权限校验的幂等重试（#811）；不可用任务的恢复有了边界，超时投递会落到 SCAN_FAILED 而不是永久 pending（#809）。
• 🧭 Skill 详情页新增"安装到 Agent"提示，完整保留 namespace 坐标（#788）；精选 Starter 集合新增固定版本的 HOL Guard plugin-scanner Skill —— 它是一个 Skill，不是 Server 扫描流水线的扩展（#814）。
• ⚙️ CLI 0.1.11：支持有边界且不破坏源文件的 Skill 升级、指纹安全的安装与同步、显式选择 namespace 拉取，以及准确的异步提交状态（#796、#804、#815、#816）。
• 🐛 正确性方面：同一坐标并发发布返回确定性冲突（#784）；并发 OAuth 身份绑定收敛，治理变更与审计记录原子化（#786、#787）；推广审核队列分页（#774）；发布后重新校验 SPA 入口文档（#812）。

🤖 Astron Agent：安全加固周期，成果发布为 v1.1.2

Astron Agent 本周期合并 30 个 PR，其中绝大部分汇入 9 月 7 日发布的 v1.1.2 安全版本（窗口结束后一天）。

• 🚨 工作流代码执行与租户隔离：<= v1.1.1 且仍使用旧版无沙箱 local 代码执行器的部署中，拥有工作流创建权限的已认证用户可在工作流服务容器内执行代码，并在受影响部署中触及其他租户数据。v1.1.2 拒绝不安全的本地执行，默认改用内置隔离的 LangChain/Pyodide 执行器，并在保留显式进程级覆盖（含有意禁用代码节点）的前提下迁移未改动的旧配置（#1650、#1651）。安全公告：GHSA-mh3w-4q3f-2fg5。感谢 Noy Pearl 与 Lasso Security 团队负责任地报告并协同复测。
• 🔒 修复插件出站请求的 SSRF（#1669），加固 XSS、制品与沙箱边界并收敛 CSRF 例外（#1643）。
• 🐳 Docker 层强制安全部署契约（#1633）、保留 MinIO 策略兼容性（#1634）、Console 使用静态迁移语句（#1640）。
• 🧩 工作流运行时修复：默认启用隔离代码回退（#1649），支持零参数代码入口（#1636）。
• 📚 文档与构建稳定性：新增"基本信息缺失导致工作流发布失败"的 FAQ（#1630）；README 增加 Related Projects 区块链接兄弟仓库（#1627）；修复日文 README 图片路径（#1653）与 VitePress 示例库构建（#1655）；补充 PyPI 源与 Maven Central 限流重试（#1660、#1657）。

⚠️ 升级提示：CODE_EXEC_TYPE=local 不再支持。请同时更新部署文件与服务镜像 —— 只更新源码检出或只更新 workflow 镜像不足以完整应用修复。

⚙️ Astron RPA：日志脱敏与 MCP 发现链路修复

Astron RPA 本周期合并 2 个 PR，均为已上报缺陷的修复。

• 🔐 敏感信息可能进入 OpenAPI、rpa-auth 与 nginx 日志（#858，关闭 #855）。新增统一脱敏器，对 API Key/Secret、bcrypt 哈希、Authorization/Token/Cookie/Session、密码、验证码、临时凭证与手机号在应用日志、Uvicorn access/error 日志、Logback 输出与异常堆栈中统一脱敏。nginx access log 改用不含查询参数的 $uri 并移除 Referer，同时保留方法、状态码、响应大小、User-Agent 与耗时等排障字段。
• 🔗 已发布工作流无法被 MCP tools/list 发现（#860，关闭 #859）。robot-service 同步工作流时使用了官方 Compose 中并不存在的容器地址，且缺失活动状态与名称等元数据。同步地址现可通过 OPENAPI_WORKFLOWS_UPSERT_URL 配置（旧地址保留为兼容回退），已发布工作流 DTO 也补齐 project_id、名称、描述、版本、status = 1 与参数。

🐬 Dolphin MCP Pilot v0.3.1：进入官方 MCP Registry

Dolphin MCP Pilot 合并 12 个 PR，并于 8 月 31 日发布 v0.3.1。

• 📇 发布到 官方 MCP Registry，标识为 io.github.iflytek/dolphin-mcp-pilot，MCP Host 与 AI 助手无需手动指向仓库即可发现并安装（#20）。发布任务依赖镜像推送完成，使用 GitHub OIDC 鉴权，先按声明的 schema 校验 server.json，并复核 GHCR 镜像的归属标签后才发布。
• 🗂️ 社区案例库：cases/ 目录收录 6 个通过验收的案例，覆盖发布闸门恢复、工作日调度安全变更、人工闸门保护下的项目清理、逐请求团队隔离与夜间 ETL 救火（#24、#39、#40、#41、#46）。
• 📦 新增 PyPI Trusted Publishing（#44）；在 PyPI 包正式可用之前，Docker 与源码仍是受支持的安装方式。

🏠 Domux：HER Hack-Astron #4 案例沉淀

Domux 合并 5 个 PR：带自动校验的案例框架（#15）与 4 个社区案例（#23、#24、#26、#29）。详见下文社区聚焦。

---

💬 社区答疑与讨论（本期精选）

以下主题提炼自 8 月 24 日—9 月 6 日的企微支持与开源讨论群匿名记录。自动问答 bot 的回答不作为权威结论 —— 本周期内多次出现错误答案，由维护者或其他社区成员在群内当场纠正。🟢 已解决 · ℹ️ 已澄清 · 🟡 需继续跟进。

Spark-X2.5

• ℹ️ 到底需要什么样的硬件？ 单路跑满 1M 上下文约需 32GB 显存；128K 上下文在 16GB 上没有问题。1.7B 纯 CPU 也可跑：社区实测 16GB 商务本上 BF16 权重约 2–3 tokens/s；量化后在 8 核 CPU 上可达 prefill 约 24 tokens/s、decode 9.5 tokens/s。非 NVIDIA 硬件方面，昇腾与海光同样支持。
• 🟢 Ollama 与 LM Studio 当前需要本地编译。在上游支持合入之前，通过 Ollama / LM Studio 运行 Spark-X2.5 需要先编译 XHToken 的 llama.cpp fork 并替换运行时 —— 群内一位 macOS 用户 LM Studio 加载报错，正是按此方式解决的。社区成员分享了可用的编译路径，也有人指出可以直接 ollama run SparkLLM/Spark-X2.5-4B，省去手动下载 GGUF 权重。随后 Spark2_5ForCausalLM 原生支持已于 9 月 6 日合入 llama.cpp 上游 b10829。

• ℹ️ 和已装的 Ollama 会冲突吗？ 先彻底关闭原生 Ollama 服务；替换后的构建同样监听 11434 端口，其他应用可继续通过原有端点调用。
• ℹ️ 4B 有 GGUF 版本吗？ 有 —— XHToken/Spark-X2.5-4B-GGUF。问答 bot 当时回答"仅 1.7B 提供 GGUF"，由社区成员在群内给出了正确链接。FP8 与 INT8 权重随后于 9 月 4 日上线，对应量化的 GGUF 预计下周发布。
• ℹ️ 4bit 量化损失有多大？ 一位贡献者在自有指令跟随任务集上对比 IQ4_NL、IQ4_XS、XS-UD、Q4_K_M 与 BF16 基准，测得 94.5%–95.8% 的能力保留率；其 imatrix 是针对该特定任务校准的，属于任务定向结论，不能直接外推为通用结论。

• 🟡 1.7B 在部分 prompt 上会出现循环重复，4B 明显更稳定；这一现象由多位社区成员与团队分别观察到。另外，文档中引用的 SGLang nightly 镜像未启用混合 SWA 池 —— 修复已提交到仓库，但在讨论发生时尚未进入已发布镜像。
• ℹ️ 有多模态/识图能力吗？ 没有，Spark-X2.5 为纯文本模型。
• ℹ️ 国内下载：推荐 魔搭 ModelScope 镜像；AtomGit、Modelers 与 SCNet 同为官方渠道。

SkillHub

• 🟢 Skill 推广到 @global 后为什么下载按钮是灰的？ 8 月 27 日由社区成员反馈，并指出该行为与官方文档"@global + PUBLIC 允许匿名搜索、查看与下载"的说明相矛盾。维护者确认这是 Bug 而非产品设计，请其提交 issue；#763 已于 9 月 2 日关闭，并在 #792 中补充回归测试。

• 🟢 推广审核页的历史待审申请"消失"了。超级管理员推广审核队列没有分页，待审申请较多时早期记录无法访问。8 月 28 日反馈为 #772，8 月 31 日关闭，分页能力随 v0.2.19 发布（#774）。
• 🟢 过期 SESSION Cookie 导致提示 "Session expired" 且必须手动清 Cookie 才能重新登录（#802，9 月 3 日关闭）。修复让浏览器携带过期会话时公共路由依然可访问（#783），并配合更大范围的会话边界加固（#801）。有成员追问 #801 是否已进入最新 release —— 它随 9 月 4 日的 v0.2.19 发布。
• ℹ️ LDAP 支持什么时候上？ 有贡献者指出：#672 看起来已合并，但路线图上仍未勾选。维护者澄清该 PR 合入的是功能分支而非 main；此前社区提交的实现通用性不足，统一身份标识方案需要重新设计后再发布。
• ℹ️ SkillHub 后续还会兼容 ClawHub CLI 吗？ 不会。持续追赶 ClawHub 每次版本更新并不可持续，建议直接使用 npm 上的 SkillHub CLI。v0.2.19 文档说明了哪些 ClawHub 的搜索、查看、安装与发布流程仍然兼容（#806）。
• 🟡 skill-scanner 显示"扫描失败 / 0 项发现"。v0.2.19 让终态 SCAN_FAILED 版本可重试（#811），并为不可用任务的恢复设定边界（#809），该状态不再是死路。但根因仍与具体部署有关：上报前请保留扫描 worker 日志与该版本的审计记录。

Astron RPA

• 🟢 MCP 调用返回 code: O000、"运行成功"，但 data 为空 {}。已于 9 月 3 日由 #860 修复。根因是已发布工作流并未真正注册到 OpenAPI 服务，因此 MCP 侧可能"成功"但工作流既不可发现也不返回输出。请升级到包含该修复的版本；若部署未使用官方 Compose 服务名，请配置 OPENAPI_WORKFLOWS_UPSERT_URL。
• ℹ️ EXE 装好后打开报 502。桌面客户端不是独立应用，必须能连通后台服务；502 表示客户端已启动但没有可用的服务端。另外请注意，GitHub 发布产物与官网发布的客户端并不等同。
• ℹ️ 开源版有 Web 控制台吗？ 没有。开源栈以桌面客户端为主，目前面向 Windows 10/11。
• ℹ️ 支持定时执行吗？ 支持 —— 直接运行、cron 表达式计划任务、调度模式与 API 调用均可用。
• ℹ️ 能对接指纹浏览器做多账号场景吗？ 目前不支持。浏览器自动化基于面向 Chrome、Edge、Firefox 的浏览器插件架构，没有公开的指纹浏览器接入或 Chrome DevTools Protocol 绑定方式。
• ℹ️ 有人大金仓（KingbaseES）的 Docker 镜像吗？ 没有提供。服务端部署的文档默认数据库为 MySQL + Redis。
• 🟡 夜间无人操作，执行记录却连续显示"用户主动关闭"。记录中未能确认根因。上报前建议保留失败时间点附近的客户端与调度日志、会话/空闲超时配置，以及主机是否会在夜间休眠、锁屏或断网 —— "用户主动关闭"应当作为现象标签，而不是结论。

Astron Agent

• ℹ️ 工作流发布失败但没有明确原因：最常见原因是工作流基本信息未填写完整，本周期已补充对应 FAQ（#1630）。
• 🚨 自建部署请尽快升级：参见上文 v1.1.2 说明。自动化漏洞扫描器未必能识别这一应用层问题，请显式核对已部署镜像版本，不要仅依赖扫描结果。

---

🔭 社区聚焦：HER Hack-Astron #4 与 #5

HER Hack-Astron #4 · Domux 智能家居真实使用案例（9 月 2 日评选完成）

参赛者需对 Domux 完成一次真实的推理、评测、量化、训练或集成实验，将完整案例发布到模型官方 Hugging Face Discussions，再通过 PR 收录进仓库 cases/ 目录。

🏆 优胜奖 —— domux-seniorsafe（Discussion #7、PR #29，@qrx-joe）。它胜出不是因为分数最好看，恰恰相反：这是本期唯一完成"先冻结、后运行、再如实公开不理想结果"三轮递进验证的案例。第二轮修复归一化拼接缺陷后，作者没有只挑成功样例，而是重跑全量数据，把 normalized exact match 从 75.7% 提升到 85.7%；同时用 raw 管线 80/80 条输出与首轮逐字节一致，反证两次实验确实使用同一模型与同一配置。第三轮先冻结 160 条挑战集、评测协议与 SHA-256 指纹，再完成 320 次真实 CPU 推理；面对 raw 55.1%、normalized 61.8%、5/24 输入策略错误放行、2/24 blocked-label 请求仍进入输出候选等不理想结果，仍然完整发布，并明确拒绝修补冻结后才发现的缺陷。案例还主动把"规则层安全 100%"标注为 true by construction, not a model metric，亲手拆掉了一个本可被包装成模型能力的漂亮指标。

🌟 优秀贡献奖：domux-4057-full-eval（RTX 4090D 上完成官方 4,057 条全量 BF16 与 NF4 对比，公开 97.58% / 93.57% 准确率与"省显存但不提速"的部署边界）；domux-4057-rtx3090-eval（PR #23，消费级显卡复现全量评测，并将正确率、并发延迟与单请求延迟分开报告）；domux-safety-gate（PR #24，从只看输入推进到检查全部七个输出字段，并如实保留 18/28 false interventions）；domux-clarify-commit（PR #26，96 次 BF16 推理接入真实 Home Assistant REST 状态变更，3/3 次精确提交、1/1 次在 dispatch 前拒绝真实状态漂移）；domux-mcp-server（四级身份认证、高危二次确认、审计日志与风险评分的完整集成方向探索）。

其中两个获奖案例未被合并 —— 奖项不自动豁免仓库的硬性验收条件：一个未通过账号信誉核验，另一个 DCO 未签署，且缺少其 README 复现命令所需的源码、测试与原始评测产物。

HER Hack-Astron #5 · Spark-X2.5 端侧智能体真实案例（9 月 6 日截止提交）

第五期在模型发布当天同步开赛，并特别邀请第四期全部参赛者把冻结评测、CPU/GPU 部署、失败分析与安全闸门等方法带入新赛题。

提交覆盖的硬件与方向相当广：Apple M3 上的工具调用安全边界；无独显 8 核/16G Linux 上 Q4_K_M 的真实编码、吞吐与 OOM 边界；24GB 内存笔记本 CPU 上的中文会议任务 JSON 提取；Intel Mac 上的中英泰多语言护栏（结论是"合法 JSON 依然可能不安全"）；面向破坏性自动化动作的中英双语 CPU 安全闸门；从坏 JSON 到百回可玩 Godot 原型；OpenClaw Agent 工具调用与安全评测 20/21；M3 上的 Agent 交接连续性测试；8GB AMD RX 580 + Vulkan 的端侧部署实测；以及一个仓库级审查案例 —— 通过三次固定协议复核定位 vLLM 0.23 的 parser 兼容断点，并把修复提交到上游 Spark-plugin #10。

提交已于 9 月 6 日 24:00（北京时间）截止，评审窗口为 9 月 7—13 日，因此本期报告不提前公布获奖结果。

👉 HER Hack-Astron #5 · Spark-X2.5 端侧智能体案例挑战赛

Hack-Astron x openKylin #1 · 让技能真正跑在国产操作系统上（9 月 3 日—24 日）

9 月 3 日与 openKylin 联合开赛的新挑战：参赛者需在 openKylin 桌面操作系统 2.0 及以上版本中本地部署 Astron Agent 与 SkillHub 开源版，使用 Astron Agent 原生 RPA 引擎实现技能逻辑，并通过 PR 提交到 openKylin-skills 仓库。仅使用 SaaS 云端版、未完成本地部署的作品不予接受。参考方向包括 UKUI 桌面自动化、软件包管理、社区运营自动化、WPS / LibreOffice 办公 RPA 与系统健康巡检。赛事比拼持续到 9 月 24 日，评奖 9 月 25—30 日。

👉 Hack-Astron x openKylin #1

社区热度

@SparkLLM 账号在模型发布后 12 小时内粉丝破 500；独立评测、视频与移动端集成实验也在数天内出现，例如 PocketPal AI 集成 XHToken 版 llama.cpp，适配 iOS 与 Android 本地推理。

---

🙌 一起共建

本周期适合上手的方向包括：

• 在我们尚未覆盖的硬件上跑 Spark-X2.5 —— 昇腾、海光、老一代消费级显卡或纯 CPU —— 并公开吞吐、内存与失败边界，而不是只给一个分数；
• 贡献一份使用通用校准集的可复现量化对比，与目前已公开的任务定向测量形成参照；
• 用最小 prompt 与固定 revision 复现 1.7B 的循环重复现象，并附原始输出提交 issue；
• 用真实的多审核人配置验证 SkillHub 的 namespace 工作区同步与新的审核/重新提交闭环；
• 用先前依赖宿主文件系统、子进程或网络访问的工作流，验证 Astron Agent v1.1.2 的隔离执行器，并反馈需要适配的点；
• 贡献一个脱敏的 Dolphin MCP Pilot 案例：真正改变工作流状态，并验证终态结果；
• 面向 Hack-Astron x openKylin Welcome to the iFLYTEK Open Source Community 👋 #1，在本地 Astron Agent + SkillHub 部署上端到端做出一个 openKylin 技能。

开工前建议先在相关 issue 下留言，便于维护者确认范围、避免重复工作。

---

🔗 探索 Astron 开源生态

• Astron Agent —— 企业级智能体工作流平台
• SkillHub —— 可私有化部署的 Agent Skill 注册与治理平台
• Astron RPA —— 面向 Agent 的桌面与 Web 自动化 RPA
• Dolphin MCP Pilot —— 面向 Apache DolphinScheduler 的生产级 MCP Server
• Domux —— 低延迟智能家居指令解析模型
• MemFlywheel —— Agent 原生长期记忆
• iFly-Skills · AstronClaw Tutorial · HarnessClaw
• Spark-X2.5 与 XHToken 组织 · Spark-X2.5 on Hugging Face
• 所有带 iflytek-astron 标签的仓库

统计口径说明：PR 数量基于 GitHub merged_at 时间戳，统计区间为 2026 年 8 月 24 日至 9 月 6 日，覆盖 iflytek-astron 标签下的 11 个仓库以及 XHToken 组织各仓库。版本日期采用 GitHub published_at。Hugging Face 下载量与 Like 数为 2026 年 9 月 8 日的快照，后续会持续变化。Astron Agent v1.1.2 发布于 9 月 7 日（窗口结束后一天），因其包含的改动均在窗口内合并，故在本期报告中一并说明。社区问题已匿名并做归纳；本文引用的截图均来自开源讨论群公开分享，凡包含可识别账号、主机或组织信息的图片均已排除；客户与群组身份、环境特有信息未予公开。

有问题、想法或希望分享贡献？欢迎参与社区讨论，或在相应仓库提交 Issue / PR。让我们继续为企业级 AI Agent 打造开放、可用于生产的基础设施 🚀