A scenario-driven recommendation engine that identifies the best Openclaw Skills to resolve workflow failures and implementation gaps.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install skill-feed
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install skill-feed using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is SkillFeed?
SkillFeed is a powerful discovery and recovery engine designed to enhance the reliability of AI agents by intelligently matching tasks to the most relevant Openclaw Skills. It functions as a diagnostic layer that monitors execution logs for errors, timeouts, or missing outputs, then dynamically generates sanitized search queries to find the perfect tool on ClawHub.
By bridges the gap between a high-level user goal and a technical implementation, SkillFeed ensures that agents remain productive even when encountering unknown challenges. It leverages a multi-stage ranking system to provide not just a single tool, but a comprehensive recovery plan tailored to the specific environment of the user.
SkillFeed Use Cases
- Automatically detecting and fixing command or API failures during an agent's task execution.
- Providing targeted Openclaw Skills recommendations when a user describes a goal without knowing which tool to use.
- Overcoming rate limits or authentication errors by suggesting alternative Openclaw Skills and fallback paths.
- Enhancing developer productivity by offering provider-specific execution steps for Claude Code, ChatGPT, and Gemini.
How SkillFeed Works
- Monitors the execution environment for failure signals like non-zero exit codes, HTTP 4xx/5xx errors, or exceeded retry thresholds.
- Captures and sanitizes the failure context to remove sensitive data like API keys or PII before searching for Openclaw Skills.
- Classifies the issue into specific categories such as auth, rate limit, network, or invalid parameters.
- Generates layered search queries ranging from broad capability searches to specific failure-token queries.
- Queries ClawHub to identify high-quality Openclaw Skills based on stars, update frequency, and goal relevance.
- Ranks the candidates and produces a structured recovery plan including primary recommendations and immediate next actions.
SkillFeed Setup
To deploy the SkillFeed logic within your automation environment, follow these configuration steps:
# Navigate to your local skills directory
cd path/to/your/agent/skills
# Initialize the SkillFeed configuration
# Ensure the discovery workflow references are available
cat references/discovery-workflow.md
Ensure your agent has the necessary permissions to access https://clawhub.ai/skills for live searching of Openclaw Skills. You should also configure your provider-specific templates in references/provider-adaptation.md to ensure the output matches your preferred AI interface.
SkillFeed Data Schema & Taxonomy
SkillFeed organizes diagnostic data and discovery results using a structured taxonomy to ensure accurate matching of Openclaw Skills:
| Data Type | Description | Key Examples |
|---|---|---|
| Goal Tokens | Keywords identifying the intended action | post, schedule, fetch, report |
| Platform Tokens | The target service or ecosystem | x, github, telegram, aws |
| Failure Tokens | Technical error markers used for filtering | 401, 429, timeout, permission denied |
| Sanitized Context | Cleaned logs used for query construction | task_name, error_type, platform |
| Recovery Plan | The final output sent to the agent | primary_skill, alternatives, actions |
SkillFeed Advanced Features
- Context-aware data sanitization that strips secrets and PII before searching for Openclaw Skills to maintain security.
- Provider adaptation logic that formats recovery steps specifically for Claude Code terminal commands or ChatGPT checklist styles.
- Layered query construction that moves from broad capabilities to failure-specific tokens for high-precision results.
- Anti-noise guardrails that prevent the system from repeating the same Openclaw Skills recommendations within a single conversation thread.
- Automated fallback paths that provide a plan B if the primary recommended skill fails to resolve the issue.
SKILL.md
Loading
Related Openclaw Skills
GEO Content Publisher
An orchestration engine designed to distribute GEO-optimized content across multiple platforms while maximizing visibility for AI crawlers and generative engines.
Binance Web3 OpenAPI Skill
A high-performance CLI skill for accessing Binance Web3 public market APIs, token security audits, and smart money signals using curated OpenAPI schemas.
Kickstarter Launch Prep
A comprehensive strategic planner for end-to-end crowdfunding success on platforms like Kickstarter and Indiegogo.
ClientTracker Freelancer CRM
A conversational CRM for freelancers to track clients, projects, and finances through simple natural language interactions.
OKR Assistant
An AI-driven goal management skill that transforms intentions into actionable OKRs with automated progress tracking and reflective reviews.
Hippocampus Memory Core
A deterministic external memory system for OpenClaw agents to store durable facts, retrieve prior decisions, and snapshot state using S3-Hipokamp.
Featured*
