integrate-fusion-agent
Integrates a Flows/Dune app with the Fusion built-in PAIA agent panel using @cognite/app-sdk. Use this skill whenever a developer wants to: open the agent panel from their app, send the agent a contextual message, let the agent read app state (resources), or let the agent call actions in the app. Triggers: "fusion agent", "PAIA", "agent panel", "sendAgentMessage", "sendAgentLayoutMode", "agent server", "registerAgentServer", "connectToHostApp", "agent integration", "agent sidebar", "app-sdk agent". Always use this skill instead of manually writing agent integration code — it sets up the correct lifecycle, graceful fallback, and recommended file structure.
72
88%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is an excellent skill description that clearly defines a specific integration domain, lists concrete actions a developer would want to perform, and provides comprehensive trigger terms covering both natural language and API-level references. The explicit 'Use this skill whenever...' clause with enumerated scenarios and the disambiguation guidance ('Always use this skill instead of...') make it highly effective for skill selection.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: open the agent panel, send contextual messages, let the agent read app state (resources), let the agent call actions in the app. Also mentions specific technical details like lifecycle setup, graceful fallback, and file structure. | 3 / 3 |
Completeness | Clearly answers both 'what' (integrates Flows/Dune app with Fusion PAIA agent panel using @cognite/app-sdk) and 'when' (explicit 'Use this skill whenever a developer wants to...' clause with specific scenarios, plus a list of trigger terms and a disambiguation note about preferring this skill over manual code). | 3 / 3 |
Trigger Term Quality | Excellent coverage of natural trigger terms including both user-facing terms ('fusion agent', 'PAIA', 'agent panel', 'agent sidebar') and API-level terms ('sendAgentMessage', 'registerAgentServer', 'connectToHostApp', 'app-sdk agent'). These cover the range of how developers would naturally refer to this integration. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive with a very specific niche: Cognite Fusion PAIA agent panel integration via @cognite/app-sdk. The domain-specific terminology (Flows/Dune, Fusion, PAIA, registerAgentServer) makes it extremely unlikely to conflict with other skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a high-quality, actionable skill with excellent executable code examples, clear step-by-step workflow, and valuable pitfall documentation. Its main weakness is that it's somewhat long for a single SKILL.md — the detailed agent server section with resources, actions, and lifecycle could benefit from being split into referenced files. The content is well-organized but could be slightly more concise in a few explanatory passages.
Suggestions
Consider extracting the agent server details (Step 5 — resources, actions, lifecycle) into a separate AGENT_SERVER.md reference file, keeping only a summary and link in the main SKILL.md.
Trim the 'Known pitfalls' Comlink proxy explanation — the symptom/fix format is good but the middle paragraph explaining React's useState mechanics is unnecessary for Claude.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is mostly efficient and avoids explaining basic concepts, but some sections could be tightened — e.g., the 'Dev vs. production' table restates what the hook already handles, and the 'Known pitfalls' section on Comlink proxies, while valuable, is somewhat verbose in its explanation. The 'Step 0 — Understand the app' section adds modest overhead. Overall good but not maximally lean. | 2 / 3 |
Actionability | Every capability has fully executable, copy-paste-ready TypeScript/TSX code with correct imports, types, and realistic patterns. The recommended file structure, factory patterns, and test examples are all concrete and immediately usable. Shell commands for installation are specific. | 3 / 3 |
Workflow Clarity | The skill clearly sequences steps 0-6 with explicit dependencies between them (e.g., 'All capabilities require a HostAppAPI instance'). It includes validation/fallback patterns (catch rejection, null checks, .catch() on register/unregister), a final checklist for verification, and clear guidance on when to use each capability. The lifecycle hook handles mount/unmount cleanup explicitly. | 3 / 3 |
Progressive Disclosure | The content is well-structured with clear headings and a logical progression, but it's a long monolithic file (~250 lines of substantive content). The recommended file structure (agentActions.ts, agentResources.ts, useAgentServer.ts) could have been split into separate reference files, and the known pitfalls / testing sections could be external references. No bundle files are provided to offload detail. | 2 / 3 |
Total | 10 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
allowed_tools_field | 'allowed-tools' contains unusual tool name(s) | Warning |
Total | 10 / 11 Passed | |
- Repository
- cognitedata/builder-skills
- Commit
d6af887
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.