Marin Cholakov
Community@marinvch · Burgas, Bulgaria
React Developer since 2021
Agent Skills by Marin Cholakov
Showing 22 vetted skills indexed across 1 GitHub repositories.
marinvchskill-creator
Draft, evaluate, and iterate on AI skills with structured evals.
marinvchcontext7
Fetch current library documentation via the Context7 API.
marinvchcontext-engineering-collection
Catalog Agent Skills for Context Engineering across production agent platforms.
marinvchvercel-react-best-practices
Apply Vercel performance patterns to React and Next.js code.
marinvchskill-template
Create structured Agent Skills with frontmatter and guided sections.
marinvchreasoning-trace-optimizer
Analyze MiniMax M2.1 interleaved thinking traces to diagnose agent failures and generate prompt improvements.
marinvchdigital-brain
Load relevant data sources as just-in-time context for AI agents.
marinvchbook-sft-pipeline
Convert books into structured SFT datasets and train author-style models with LoRA fine-tuning.
marinvchcomprehensive-research-agent
Enforce validation, error recovery, and transparent reasoning in multi-source research tasks.
marinvchmulti-agent-patterns
Coordinate multiple agents with supervisor, swarm, and hierarchical patterns.
marinvchbdi-mental-states
Converts RDF context into Belief-Desire-Intention mental states with justifications and temporal validity for multi-agent systems.
marinvchcontext-degradation
Detect and diagnose five context degradation patterns in long-running AI agent conversations.
marinvchcontext-compression
Compress long agent conversations into anchored structured summaries preserving session context.
marinvchmemory-systems
Design cross-session agent memory with layered vector, graph, and temporal storage.
marinvchadvanced-evaluation
Design and run LLM evaluation workflows with structured rubrics and bias mitigation.
marinvchcontext-fundamentals
Optimize AI agent context budgets with modular prompts and token estimation.
marinvchproject-development
Plan end-to-end LLM projects with pipeline and cost estimation.
marinvchevaluation
Score agent outputs against a multi-dimensional rubric with per-dimension and overall scores.
marinvchcontext-optimization
Optimizes token usage via KV-cache, masking, partitioning for limited AI contexts.
marinvchfilesystem-context
Offload tool outputs and context to the filesystem for persistent agent workflows.
marinvchtool-design
Convert human-facing APIs into structured tool descriptions with consistent schemas.
marinvchhosted-agents
Orchestrate hosted agents in sandboxed execution environments for remote user tasks.