From researcher
Orchestrates the implementation and code-analysis skills to read source code, generate pseudocode, and write reproducible experiment scripts; invoke for code-to-paper tasks.
How this agent operates — its isolation, permissions, and tool access model
Agent reference
researcher:agents/code-agentsonnetSkills preloaded into this agent's context
The summary Claude sees when deciding whether to delegate to this agent
Handles code analysis and implementation tasks. - implementation - code-analysis This agent runs on Sonnet, set via the `model: sonnet` frontmatter field above. The implementation and code-analysis skills fork into this agent automatically, each carrying its own `context: fork` and `agent: code-agent` frontmatter; there is no markdown-level enforcement here. Opus-tier reasoning stays in the mai...
Handles code analysis and implementation tasks.
This agent runs on Sonnet, set via the model: sonnet frontmatter field above. The implementation
and code-analysis skills fork into this agent automatically, each carrying its own context: fork and
agent: code-agent frontmatter; there is no markdown-level enforcement here. Opus-tier reasoning stays
in the main session for research, writing, and review tasks.
npx claudepluginhub sokolmarek/researcherOrchestrates structured AI-driven research workflows for software development, from codebase exploration and prior-art review through planning, experimentation, implementation, validation, and handoff.
Implements analysis pipelines, data processing, ML/DL training, simulations, and statistical tests. Primarily Python, with full tool access for debugging computational experiments.
Authorized local subagent for writing ephemeral experimental code. Runs isolated Docker testbeds to compare 3+ approaches, producing evidence-based recommendations and permanent RESULTS.md documentation.