Agentic Dev Io Mcp Code Execution
agentic-dev-io-mcp-code-execution is a Claude Code plugin for the Build phase that makes MCP server interaction token-efficient by executing interaction code with progressive tool disclosure.
by agentic-dev-io · github.com/agentic-dev-io/mcp-code-execution
Install when your agent burns context on MCP tool schemas and you want server calls executed in-environment with progressive disclosure instead of loading every tool into the prompt.
Add it to Claude Code
Install the plugin in Claude Code. One command, paste-ready.
/plugin install agentic-dev-io-mcp-code-execution@agentic-dev-io/mcp-code-executionBuilt to be called by your agent
Skillselion is itself an MCP server. Your agent can pull this entry and a paste-ready install config straight from the API - no copy-paste.
Retrieve this entry with skillselion.get_details("plugin:agentic-dev-io/mcp-code-execution") and the paste-ready config with skillselion.get_install_config("plugin:agentic-dev-io/mcp-code-execution").
What it does
agentic-dev-io-mcp-code-execution is a Claude Code plugin bundle aimed at solo builders who depend on multiple MCP servers and hit context limits from large tool manifests. Rather than treating every MCP tool as prompt surface area, the approach emphasizes executing interaction logic as code and revealing tools progressively so the agent only pays for what the current step needs. That fits teams shipping agent workflows, internal CLIs, and API-backed automations where MCP is the integration spine. You reach for it during Build when you are wiring servers, tuning agent sessions for long tasks, or comparing MCP strategies against raw tool-calling. It is an integration-oriented plugin, not a full agent skill library: value is in how MCP access is structured for token efficiency. Pair it with your existing MCP servers and Claude Code; expect configuration around code execution and disclosure rules rather than a turnkey SaaS UI.
Highlights
- Runs MCP server interaction through code execution instead of dumping full tool definitions into context
- Claims 95–99% reduction in context usage via progressive tool disclosure
- Processes tool results in-environment before summaries return to the agent
- Single-plugin bundle focused on token-efficient MCP patterns
- Community catalog entry from agentic-dev-io/mcp-code-execution repository
Why builders use it
Loading every MCP tool and result into the agent context makes long coding sessions expensive and brittle.
After install, MCP calls can run through in-environment execution with staged tool exposure so context stays small while integrations keep working.
At a glance
- Type - Plugin in LLM Integration.
- Adoption - 0 installs, 1 stars, 0 votes.
FAQ
Who is agentic-dev-io-mcp-code-execution for?
It is for Claude Code users building agent workflows that talk to MCP servers and want materially lower context use during those sessions.
When should I use agentic-dev-io-mcp-code-execution?
Use it while integrating MCP servers during Build, especially when tool lists are large or sessions chain many server calls.
How do I add agentic-dev-io-mcp-code-execution to my agent?
Register the plugin from the agentic-dev-io/mcp-code-execution repository in Claude Code’s plugin flow, then point your MCP config at servers that support the execution and disclosure model.
Comments
Share how you use agentic-dev-io-mcp-code-execution, gotchas, or tips for other indie builders.
No comments yet - be the first to share how you use it.