Appxiom MCP Skills Installation
Package details and releases are available on npm at @appxiom/mcp-skills.
The Appxiom MCP Skills package provides pre-packaged AI Agent Skills and runbooks for teams using the Appxiom Model Context Protocol (MCP) server. These skills teach AI coding assistants (such as Antigravity IDE, Claude Code, GitHub Copilot, Cursor, Windsurf, and Xcode) standard diagnostic and prioritization protocols for real-time issues, goal conversion drop-offs, app quality scores, and deep occurrence telemetry.
Included Skills
| Skill | Installed Directory | Description | Primary MCP Tools |
|---|---|---|---|
goals | appxiom-goal-friction-analysis/ | Multi-factor ranking of goal conversion drop-offs, device impacts, and funnel health to prioritize bugs with the highest business impact | get_gfi_goals, get_gfi_issues, get_issue_details |
severity | appxiom-issue-severity-classification/ | Standardized workflow to filter and prioritize issues by severity level (fatal, major, minor) based on crash rates and user scope | list_issues, get_issue_details, get_issue_analytics |
quality | appxiom-quality-score-interpretation/ | App Quality Score (0–10) metrics evaluation, health band mapping (Healthy, Stable, Warning, Critical), and release stability breakdown | get_quality_score |
telemetry | appxiom-occurrence-telemetry-analysis/ | Root-cause investigation across user activity trails, function parameters, network calls, device specs, and hardware runtime states | list_issue_occurrences, get_issue_occurrence_details |
Installation via NPX
Use the npx @appxiom/mcp-skills add command to install skills directly into your project.
1. Install All Skills (Recommended)
To install the complete suite of Appxiom MCP skills:
npx @appxiom/mcp-skills add all
2. Install Individual Skills
You can also install individual skills as needed:
# Install Goal Friction Analysis skill
npx @appxiom/mcp-skills add goals
# Install Issue Severity Classification skill
npx @appxiom/mcp-skills add severity
# Install Quality Score Interpretation skill
npx @appxiom/mcp-skills add quality
# Install Occurrence Telemetry Analysis skill
npx @appxiom/mcp-skills add telemetry
Target Directory Detection
The CLI automatically detects your project's AI assistant configuration and places skills in the appropriate directory:
- If
.claude/exists → installs to.claude/skills/<skill-name>/ - If
.agents/exists → installs to.agents/skills/<skill-name>/ - Default fallback → installs to
.agents/skills/<skill-name>/(supported by Antigravity, Cursor, and agentic workflows)
Specifying Target Agent
You can explicitly target a specific AI agent environment using the --agent option:
# Target Claude Code / Claude Desktop (.claude/skills)
npx @appxiom/mcp-skills add all --agent claude
# Target Antigravity IDE & Agentic workspaces (.agents/skills)
npx @appxiom/mcp-skills add all --agent agents
# Target GitHub Copilot (.agent/skills)
npx @appxiom/mcp-skills add all --agent copilot
Custom Target Directory
To install into a custom or editor-specific directory (such as Cursor or a custom workspace path), use the --target flag:
# Install into Cursor skills directory
npx @appxiom/mcp-skills add all --target .cursor/skills
# Install into a custom directory
npx @appxiom/mcp-skills add all --target ./my-custom-skills-dir
Skill Capabilities & Example Prompts
Once installed, your AI coding agent can automatically leverage these skills when responding to prompts:
1. Goal Friction Analysis (appxiom-goal-friction-analysis)
Teaches the agent to read GFI scores, drop-off percentages, affected device counts, and conversion funnel health together.
- "Which checkout bugs should we fix first based on goal conversion impact?"
- "Evaluate the GFI score and drop-off rate for our onboarding flow."
2. Issue Severity Classification (appxiom-issue-severity-classification)
Enables standardized triage and investigation across severity levels.
- "List all fatal crashes affecting the latest release and show their stack traces."
- "Show all major non-fatal issues impacting more than 5% of our active users."
3. Quality Score Interpretation (appxiom-quality-score-interpretation)
Translates raw 0–10 Quality Scores into actionable health insights and penalty breakdowns.
- "What is our overall Quality Score for the latest release, and why did it drop?"
- "Explain the severity drivers impacting our release stability."
4. Occurrence Telemetry Analysis (appxiom-occurrence-telemetry-analysis)
Provides the agent with the schema and field definitions for user breadcrumbs, function arguments, and device state.
- "Analyze the activity trail leading up to crash ID 4819."
- "Inspect the memory state, device battery, and function parameters for the latest occurrence of issue 102."