constellation

Get a high-level overview of the codebase architecture

Analyze dependencies for a file or symbol

Quick health check for Constellation connectivity and authentication

Analyze the impact of changing a symbol or file

Check Constellation API connectivity and authentication status

Find orphaned/dead code that is exported but never imported

Use this agent proactively for ANY dependency-related task - checking imports, understanding module relationships, finding circular dependencies, or analyzing coupling. Trigger this agent whenever imports, dependencies, or module structure is discussed. Examples: <example> Context: User is adding new imports to a file user: "I need to import UserService into AuthService" assistant: "Let me check if adding this import would create a circular dependency or unhealthy coupling." <commentary> Adding imports between services can create circular dependencies. The dependency-detective should verify this import won't create cycles. </commentary> </example> <example> Context: User is reviewing the codebase architecture user: "I'm worried our modules are too tightly coupled" assistant: "I'll analyze the module dependencies to identify any circular dependencies and tightly coupled areas." <commentary> Architecture concerns about coupling warrant a dependency analysis. The agent should map out module relationships and identify problems. </commentary> </example> <example> Context: User is preparing for a code review or PR user: "Before I submit this PR, are there any dependency issues I should know about?" assistant: "I'll scan for circular dependencies and check if your changes introduced any new coupling issues." <commentary> Pre-PR checks are a good time to verify dependency health. The agent should check for cycles and report any concerns. </commentary> </example>

Use this agent proactively before ANY code modification to assess impact. Trigger whenever the user discusses changing, refactoring, renaming, deleting, or modifying code - even for seemingly small changes. Always analyze before acting. Examples: <example> Context: User has been discussing refactoring a core service class user: "I'm going to rename UserAuthService to AuthenticationService" assistant: "Before renaming UserAuthService, let me analyze the potential impact of this change to understand the blast radius." <commentary> The user is about to make a significant change to a service class. The impact-investigator should be triggered to assess which files and symbols depend on this class before proceeding. </commentary> </example> <example> Context: User is working on deleting unused code user: "I think this validateInput function is dead code, going to remove it" assistant: "Let me verify that validateInput is truly unused and check for any hidden dependencies before removal." <commentary> Before deleting code, it's prudent to verify it's actually unused. The impact-investigator can confirm this and identify any places that might dynamically reference the function. </commentary> </example> <example> Context: User is planning a major refactoring of a module user: "We need to restructure the payment processing module" assistant: "I'll analyze the payment module to understand its dependencies and dependents, helping identify the safest order for these changes." <commentary> Major restructuring benefits from understanding the full dependency graph. The agent should map out what depends on what to suggest a safe refactoring order. </commentary> </example>

Use this agent for any codebase exploration, understanding, or navigation task. This agent proactively uses Constellation to provide intelligent code search, symbol discovery, and architectural understanding. Examples: <example> Context: User wants to understand the codebase user: "What does this codebase do?" assistant: "I'll use Constellation to analyze the architecture and give you an overview." <commentary> Open-ended codebase questions should trigger intelligent analysis using getArchitectureOverview. </commentary> </example> <example> Context: User is looking for where something is implemented user: "Where is the authentication logic?" assistant: "I'll search for authentication-related symbols and trace their implementations." <commentary> Implementation questions should use searchSymbols and traceSymbolUsage instead of grep. </commentary> </example> <example> Context: User wants to find usages of something user: "Find all places that use the UserService" assistant: "I'll trace all usages of UserService across the codebase." <commentary> Usage questions should use traceSymbolUsage for accurate results. </commentary> </example> <example> Context: User asks how something works user: "How does the payment processing work?" assistant: "I'll analyze the payment-related symbols and their call graphs to explain the flow." <commentary> "How does X work" questions benefit from getCallGraph and dependency analysis. </commentary> </example> <example> Context: User wants to explore a specific area user: "Show me the API endpoints in this project" assistant: "I'll search for controller and route definitions using Constellation." <commentary> Exploration of specific code areas should use targeted symbol search. </commentary> </example>

This skill should be used when the user asks about "Constellation errors", "fix Constellation", "debug Constellation", "Constellation not working", "API connection issues", "indexing problems", "MCP server", "Failed to reconnect", mentions any Constellation error codes (AUTH_ERROR, PROJECT_NOT_INDEXED, etc.), or when Constellation commands fail.

4 hooks across 4 events