agent-specific-mcp-server-initialization

agent-specific-mcp-server-initialization is a skill for Claude Code, Codex from ychampion/cskill-agents. It costs 32 tokens per session (405 once invoked), scanned A, original, MIT.

A procedure for connecting extra MCP servers declared by an agent and cleaning them up when that agent finishes. MCP servers are external services that provide tools or data to a coding agent.

In plain words
What is it for?
Use it when child agents declare additional MCP servers, so their named or inline connections can be added to the session and agent-owned connections can be removed safely afterward.
Why use it?
It prevents an agent from losing the tools it declares and avoids closing shared connections that other agents still need.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/ychampion/cskill-agents/agent-specific-mcp-server-initialization
Any agent
npx skills add ychampion/cskill-agents --skill agent-specific-mcp-server-initialization
Clone the repo
git clone --depth 1 https://github.com/ychampion/cskill-agents

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for agent-specific-mcp-server-initialization

README.md
[![agentmods](https://agentmods.dev/badge/skills/ychampion/cskill-agents/agent-specific-mcp-server-initialization.svg)](https://agentmods.dev/skills/ychampion/cskill-agents/agent-specific-mcp-server-initialization)
Your own site
<a href="https://agentmods.dev/skills/ychampion/cskill-agents/agent-specific-mcp-server-initialization"><img src="https://agentmods.dev/badge/skills/ychampion/cskill-agents/agent-specific-mcp-server-initialization.svg" alt="Measured on agentmods" height="20"></a>
Per session 32 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 405 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00032 $0.00405
Opus 5 $0.00016 $0.00202
Sonnet 5 $0.00006 $0.00081
Haiku 4.5 $0.00003 $0.00040

Measured 4d ago against content hash 04e677cc978a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agent-specific-mcp-server-initialization scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/claude-code/skills/agent-specific-mcp-server-initialization/SKILL.md · 28 lines

What it actually says

SKILL: Agent-Specific MCP Server Initialization

Domain: mcp-initialization Trigger: Load this skill when an agent defines extra MCP servers in its frontmatter and those clients must be joined to the parent session. Source Pattern: Distilled from reviewed subagent orchestration, isolation, and lifecycle implementations.

Core Method

Iterate the servers declared by the child agent, resolve named references from shared MCP configuration, and treat inline declarations as new agent-owned connections. Merge the resulting clients and tools into the inherited parent context, but keep track of which connections were created specifically for that child agent. When the agent ends, tear down only those agent-owned connections while leaving inherited shared clients intact. If no extra servers are declared, return the parent context unchanged.

Key Rules

  • Short-circuit when the agent declares no extra MCP servers and return the inherited clients unchanged.
  • For string specs, reuse memoized MCP configs; for inline objects, mark the resulting clients as newly created so their cleanup runs after the agent ends.
  • Always return the merged client list plus the fetched tools, and expose a cleanup closure that only targets the newly created clients.
  • Guard the loop with logging for missing configs so failures degrade gracefully instead of throwing.

Example Application

When implementing a custom agent that needs access to an inline MCP server, plug into this skill to merge the new server with the parent clients, reuse any shared configs, and ensure only the inline connection runs its cleanup path.

Anti-Patterns (What NOT to do)

  • Don’t rebuild the entire MCP client list from scratch every turn; reuse the inherited parent clients and append the new ones instead.
  • Don’t forget to clean up inline MCP servers; leaking them leaves dangling connections and stale credentials.
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago First seen · 28 lines · 32 tokens per session scan A 04e677cc978a

Subscribe to this mod's changes

agent-specific-mcp-server-initialization is a skill published in the GitHub repository ychampion/cskill-agents (36 stars, last pushed 5mo ago), licensed MIT. It adds 32 tokens to every session and 405 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

rpce-merge-pr-batch

Safely process an explicitly ordered batch of RepoPrompt CE pull requests end to end: preserve a dirty original checkout, isolate each PR in an external disposable worktree, use window- and context-scoped CE rpce-cli Agent Mode review, repair and validate, require exact-head hosted checks, merge with merge commits…

repoprompt/repoprompt-ce · 128 tokens

rpce-contribution-check

Validate RepoPrompt CE contributions before committing or pushing. Use whenever an agent is about to create a commit, push the current branch, rewrite history, delete a branch or fork, or change GitHub-visible repository state. Enforces staged-index and outgoing-range secret scanning, repository guardrails, clean push…

repoprompt/repoprompt-ce · 90 tokens

rpce-test-quality

Select, design, review, consolidate, or remove RepoPrompt CE tests, diagnostic harnesses, and smoke checks by regression value and maintenance cost. Use when the task centers on test, diagnostic, or smoke coverage, including whether a single regression test is worth committing. Do not use for feature or bug-fix work…

repoprompt/repoprompt-ce · 83 tokens

rpce-release

Build or publish RepoPrompt CE release artifacts using the repository release scripts and GitHub workflows.

repoprompt/repoprompt-ce · 22 tokens

rpce-maintainer-guidance

Apply evidence-led RepoPrompt CE maintainership principles distilled from documented project guidance. Use when planning, scoping, implementing, reviewing, triaging, or sequencing RPCE changes; deciding whether work should be an investigation, issue, PR, or follow-up; checking compatibility, migrations, performance…

repoprompt/repoprompt-ce · 87 tokens

rpce-swift-6-concurrency-migration

Plan, inventory, stage, execute, or review RepoPrompt CE's project-wide migration to Swift 6.2 concurrency checking and Swift 6 language mode. Use when auditing packages, targets, settings, diagnostics, unsafe escape hatches, migration phases, blockers, or validation evidence across the root and provider packages. Do…

repoprompt/repoprompt-ce · 90 tokens