codescope

End-to-end MCP server verification — spawn stdio server, list tools, invoke each one, verify response shape.

Set up a demo session — pick a sample repo, index it, open web UI, prep the narrative.

Run full lint + format + test + frontend build. Pre-commit / pre-release check.

Cut a release — bump version, update CHANGELOG, tag, push, create GitHub release, install local binaries.

Sprint / project status from knowledge graph. Show what's planned, in-progress, done.

Audit MCP tool set — count, descriptions, consolidation candidates, budget compliance.

Autonomous research loop. Takes a topic, runs web searches, fetches sources, synthesizes findings, and files everything into the knowledge graph. Based on Karpathy's autoresearch pattern. Triggers on: "/autoresearch", "research [topic]", "deep dive into [topic]", "investigate [topic]", "find everything about [topic]", "go research".

Ingest any source (file, URL, image) into the codescope knowledge graph. Extracts entities, concepts, claims, and cross-references them with existing code entities. Triggers on: "ingest this", "add to knowledge base", "read and file", "process this source", "ingest [url]", "wiki ingest".

Health check the knowledge graph. Finds orphan nodes, missing cross-references, stale claims, unresolved contradictions, and knowledge gaps. Triggers on: "lint the wiki", "knowledge health", "check knowledge base", "wiki lint", "find orphans".

Query the knowledge graph for answers. Searches knowledge entities, claims, sources, and cross-references with code. Cites specific graph nodes, not training data. Triggers on: "what do we know about", "query knowledge", "check the wiki", "knowledge search".

Code intelligence engine — query knowledge graphs, find functions, trace callers, analyze impact. Use when user asks about code structure, dependencies, function relationships, or wants to understand a codebase.

Ask a natural language question about the codebase. Translates to graph queries automatically. Use when user asks questions in Turkish or English about code structure.

Find all callers of a function. Use when user asks who calls a function, wants to trace call chains, or asks about function usage.

List all entities (functions, classes, imports, configs) in a specific file. Use when user asks what's in a file, wants a file overview, or asks about file contents.

Analyze the impact of changing a function. Shows callers, dependencies, and affected files. Use when user asks what happens if they change or refactor something.

Re-index the current codebase with Codescope. Use when the user wants to refresh the knowledge graph after code changes.

Execute a raw SurrealQL query on the knowledge graph. Use for advanced queries, custom reports, or when other tools are not specific enough.

Search functions in the Codescope knowledge graph. Use when user wants to find functions, methods, or symbols by name pattern.

Show codebase statistics from the Codescope knowledge graph. Use when user asks how many functions, classes, files, or wants a project overview.

Check for Codescope updates and install the latest version. Self-update mechanism.

Code intelligence engine — query knowledge graphs, find functions, trace callers, analyze impact. Use when user asks about code structure, dependencies, function relationships, or wants to understand a codebase.

Execute a raw SurrealQL query on the knowledge graph. Use for advanced queries, custom reports, or when other tools are not specific enough.