← Back to Skills Marketplace
Meta Skill Weaver
by
pagoda111king
· GitHub ↗
· v2.5.0
· MIT-0
270
Downloads
0
Stars
1
Active Installs
8
Versions
Install in OpenClaw
/install meta-skill-weaver
Description
Use this skill when orchestrating complex multi-step tasks. Provides first-principles task decomposition, EventBus event system, multi-skill collaboration wi...
Usage Guidance
This skill appears to implement the orchestration features it advertises, but there are several unexplained inconsistencies and some runtime vectors that need clarification before installation. Before you install or enable it broadly: 1) Ask the maintainer for the canonical source repository and a signed/consistent release (the metadata/contacts/versions in the package are inconsistent). 2) Request a clear install spec (what the installer does, exact npm dependencies from package.json) and run npm audit on the package-lock in a sandbox. 3) Confirm where metrics/feedback and 'virtual path' state are persisted (local sandboxed storage vs remote endpoints) and whether any network endpoints are contacted; get explicit env vars required for remote storage if any. 4) Verify what arbitrary handlers (e.g., notify.sh) are allowed to run and ensure they cannot be supplied by untrusted users; avoid enabling arbitrary shell execution handlers in production. 5) If you must test it, run it in an isolated sandbox/container with no access to production secrets, and review src/middleware/*.js for any outbound network calls or uses of child_process/exec. If the maintainer cannot answer these questions or provide a single authoritative repository and release notes, treat the package as untrusted.
Capability Analysis
Type: OpenClaw Skill
Name: meta-skill-weaver
Version: 2.5.0
The bundle is classified as suspicious due to significant anomalies in the `package-lock.json` and extreme inconsistencies across documentation. The dependency tree has been modified to inject non-standard packages like `unrs-resolver` and `exit-x` into core Jest components; `unrs-resolver` is particularly concerning as it includes an install script and native bindings for multiple platforms, a common pattern in supply chain attacks. Furthermore, the bundle references a non-existent or unreleased version of Jest (30.3.0) and contains conflicting version numbers and support emails across `SKILL.md`, `README.md`, and `package.json`. While the provided middleware source code in `src/` appears benign, the compromised state of the dependency lockfile poses a high risk if the agent attempts to install or test the package.
Capability Assessment
Purpose & Capability
The name/description claim an orchestration/meta-skill (task decomposition, EventBus, middleware, metrics, feedback). The bundled files (middleware implementations, tests, coverage, package.json/lock) are consistent with that purpose. However, the package includes source and a package-lock but the skill registry entry declares no install spec; SKILL.md also references a 'clawhub install' CLI command — this mismatch (packaged source but no declared install mechanism) is unexpected and should be clarified. There are also inconsistent maintainer/contact strings across files (pagoda111king, cloud-shrimp, '王的奴隶'), and differing version numbers in headers, which reduce confidence in provenance.
Instruction Scope
SKILL.md instructs the agent to run CLI commands (meta-skill-weaver start/status/resume/cancel) and to register event handlers (e.g., meta-skill-weaver on 'task.complete' --handler="notify.sh"). Allowing arbitrary shell handlers (notify.sh) and Exec/Bash tools (declared in SKILL.md) means runtime can execute arbitrary scripts on the host when used. The documentation discusses persisting metrics/feedback and 'virtual path' storage but does not specify storage locations or whether network endpoints are contacted; that ambiguity could allow data exfiltration if middleware is configured to send metrics or feedback off-host. The instructions do not read unrelated system files or request credentials explicitly, but their open-ended handler execution and unspecified persistence endpoints increase risk.
Install Mechanism
No install spec is declared in the registry (instruction-only), but the bundle contains package.json and a large package-lock, source JS files, and coverage artifacts — implying an npm project that expects to be installed. That mismatch is notable: a consumer following SKILL.md's 'clawhub install meta-skill-weaver' may download/execute code not described by the registry entry. Lack of an explicit, reviewable install mechanism increases friction for safe vetting (you can't tell from metadata which packages would be installed).
Credentials
The registry declares no required environment variables or credentials, which is consistent with SKILL.md. At the same time, SKILL.md mentions built-in middleware including 'Auth' and persistent metrics/feedback storage, which plausibly require external credentials or storage config — none are declared. This absence of declared env/config requirements while describing features that typically need storage/auth is an information gap that should be resolved before trusting the skill with sensitive data.
Persistence & Privilege
The skill does not request forced persistence (always:false) and uses normal autonomous invocation settings (disable-model-invocation:false), which is expected. It documents saving task state to a 'virtual path' and persisting metrics/feedback, but does not state where those artifacts live (local filesystem, external DB, cloud). Because the skill can persist state and execute shell handlers, confirm storage locations and access controls; though not an outright privilege escalation, the combination of persisted state + arbitrary handler execution increases the blast radius if the code were malicious or misconfigured.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install meta-skill-weaver - After installation, invoke the skill by name or use
/meta-skill-weaver - Provide required inputs per the skill's parameter spec and get structured output
Version History
v2.5.0
v2.5.0: Added MetricsMiddleware and FeedbackMiddleware. All 54 tests pass.
v2.4.0
v2.4.0: MetricsMiddleware auto-optimization + FeedbackMiddleware user feedback + 54 new tests
v0.3.0
v0.3.0: 增强任务分解、技能分配、结果聚合能力
v0.4.1
O 维度提升 - 添加事件钩子系统,支持技能编排链式调用
v0.4.0
meta-skill-weaver v0.4.0 – Major "state persistence" upgrade
- Added State Machine system for advanced task state tracking and management.
- Introduced State Persistence, enabling JSON-based state backups, rollback, and robust interrupt recovery.
- Added State Persistence Middleware for seamless integration of state saving/restoring in workflow chains.
- 31 new test cases (total: 92); 73.72% statement coverage.
- Upgraded multi-tasking, parallel execution, and recovery: now lists all recoverable tasks with full state history.
- Documentation, module layout, and six-dimension scorecard updated to reflect new architecture and capabilities.
v2.3.1
- Expanded and reorganized SKILL.md documentation for clarity, with more detailed usage scenarios, quickstart, architecture overview, and FAQs.
- Added explicit sections for use cases, advanced usage examples, technical architecture, file structure, success stories, and support contacts.
- Clarified parallel execution, event-driven workflow, virtual path isolation, concurrent limits, and middleware customization in FAQ and architecture diagrams.
- Included command-line examples and practical flows for real-world orchestration tasks.
- Minor version bump to 2.3.1; no changes to code or core functionality.
v2.3.0
meta-skill-weaver v2.3.0
- Major refactor: migrated documentation, removed legacy modules, updated for production stability.
- SKILL.md rewritten with concise English/Chinese overview, event-driven architecture details, feature matrix, pricing tiers, and FAQ.
- Deprecated and removed middleware-chain, parallel-execution, task-tracker, and virtual-path-system source files.
- Added 35 Jest test cases (62% coverage), enhancing reliability.
- Updated package.json and documentation structure to reflect new features and codebase simplification.
v0.1.0
meta-skill-weaver v0.1.0 changelog:
- Initial release of a task orchestration and execution engine.
- Supports complex task decomposition into executable subtasks.
- Assigns appropriate skills to each subtask based on task type.
- Tracks task progress in real-time, including status definitions (pending, running, blocked, completed, etc).
- Provides interruption recovery, allowing tasks to resume from breakpoints.
- Collects execution and optimization data for continuous skill improvement.
Metadata
Frequently Asked Questions
What is Meta Skill Weaver?
Use this skill when orchestrating complex multi-step tasks. Provides first-principles task decomposition, EventBus event system, multi-skill collaboration wi... It is an AI Agent Skill for Claude Code / OpenClaw, with 270 downloads so far.
How do I install Meta Skill Weaver?
Run "/install meta-skill-weaver" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Meta Skill Weaver free?
Yes, Meta Skill Weaver is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Meta Skill Weaver support?
Meta Skill Weaver is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Meta Skill Weaver?
It is built and maintained by pagoda111king (@pagoda111king); the current version is v2.5.0.
More Skills