Arthur Korochansky
Community@artk0de · Somewhere in space
The Developer
Agent Skills by Arthur Korochansky
Showing 35 vetted skills indexed across 1 GitHub repositories.
debug-pipeline-log
Analyze TeaRAGs pipeline debug logs to detect performance bottlenecks and timing anomalies.
add-migration
Create and register stateful migrations for TeaRAGs-MCP schema, sparse vector, and snapshot pipelines.
expand-coverage
Analyze coverage JSON reports and generate behavioral tests for uncovered branches.
test-self-reindex
Reindex a worktree source to validate the TeaRAGs MCP indexing and enrichment pipeline.
add-rerank-preset
Define custom reranking presets for MCP search tools with scoring weights and overlay masks.
add-language-hook
Define custom chunking hooks for programming languages in the TeaRAGs ingest pipeline.
optimize-skill
Optimize AI agent skill instructions through structured eval cycles and baseline verification.
add-dto
Standardize request and response DTO definitions for MCP server endpoints.
add-derived-signal
Implement custom normalized ranking signals for TeaRAGs code reranking.
add-mcp-endpoint
Create and register MCP tools following TeaRAGs-MCP DTO, facade, and registration patterns.
analytics-rerank
Reorder code search results using git-based structural signals.
risk-assessment
Analyze codebase risk by cross-referencing git history with semantic search.
extract-project-patterns
Identifies stable code templates from a project using git-derived metrics.
explore
Combine breadth-first discovery with depth-first tracing for code investigation.
Refactoring Scan
Analyze codebases to identify and prioritize refactoring candidates using churn, complexity, and stability metrics.
bug-hunt
Ranks code chunks by git bug-fix rates and churn to identify root causes.
data-driven-generation
Generate code by selecting strategies from git-derived stability and risk signals.
index
Register project aliases and perform incremental indexing via the TeaRAGs MCP server.
force-reindex
Rebuild the entire codebase search index in a background process.
Pattern Search
Group code pattern implementations using semantic search and git history.
filter-building
Translate natural language scopes into typed search filters for TeaRAGs MCP server.
tests-as-context
Enrich semantic code search results with DSL test chunks and git-based stability signals.
report-issue
Collect diagnostic data and verify duplicates before creating GitHub issues.
test-driven-development
Generate project-specific tests by searching existing test patterns and conventions.