Skip to main content
hc
Submit
skillsSource-backedReview first Safety ✓ Privacy ✓

Subagent MCP Scoping Review Capability Pack Skill

Expert subagent MCP scoping review capability pack for auditing which MCP servers and tools subagents inherit, restricting high-risk tools in delegated tasks, and preventing unintended production changes from background agents.

by kiannidev·added 2026-06-14·
Claude CodeCodexWindsurfGeminiCursorCLI
HarnessClaude CodeCodexWindsurfGeminiCursorCLI
Level:expertType:capability-packVerified:validated
Review first — review before installing

Open the source and read safety notes before installing.

Safety notes

  • Subagents operate in isolated context windows but may inherit MCP access from the parent configuration unless explicitly restricted.
  • Background subagents can invoke write or execute MCP tools without the same interactive oversight as foreground work.
  • Delegating production-impacting tasks to subagents does not reduce accountability for resulting changes.
  • Removing MCP servers from the parent session before delegation is safer than assuming subagents ignore high-risk tools.
  • This skill recommends scope reductions; it must not spawn subagents with broader MCP access than the reviewed plan allows.

Privacy notes

  • Subagent MCP reviews may expose internal service names, database tools, and deployment integrations configured in the parent session.
  • Delegated task descriptions and subagent transcripts can contain customer data returned by MCP read tools.
  • Public summaries should describe scope categories and mitigations, not live MCP manifests or tool arguments.

Prerequisites

  • An planned subagent delegation with a defined task boundary and expected outputs.
  • Inventory of MCP servers configured in the parent Claude Code session or repository.
  • Knowledge of whether the subagent will run foreground or background and who approves tool use.
  • Redacted task description sufficient to classify read versus write MCP tool needs.

Schema details

Install type
package
Reading time
9 min
Difficulty score
82
Troubleshooting
Yes
Breaking changes
No
Source repository stats
Scope
Source repo
Package metadata
Skill and platform metadata
Skill type
capability-pack
Skill level
expert
Verification
validated
Verified at
2026-06-14
Retrieval sources
https://code.claude.com/docs/en/sub-agentshttps://code.claude.com/docs/en/mcphttps://code.claude.com/docs/en/skillshttps://code.claude.com/docs/en/features-overviewhttps://github.com/anthropics/claude-codehttps://developers.google.com/search/docs/fundamentals/creating-helpful-content
Tested platforms
ClaudeClaude CodeCodexCursorWindsurfGeneric AGENTS
PlatformSupportInstall path
claude-codeNative.claude/skills/<skill-name>/SKILL.md
codexNative.agents/skills/<skill-name>/SKILL.md
windsurfNative.windsurf/skills/<skill-name>/SKILL.md
geminiNative.gemini/skills/<skill-name>/SKILL.md or .agents/skills/<skill-name>/SKILL.md
cursorAdapter.cursor/rules/<skill-name>.mdc
cliManualAGENTS.md or tool-specific context file
Full copyable content
# Trigger
"Apply the subagent MCP scoping review capability pack for this delegation."

# Required output
1) Parent session MCP inventory versus subagent inheritance summary
2) High-risk tool exposure assessment for the delegated task
3) Recommended MCP scope reduction or approval gates
4) Foreground versus background delegation safety notes
5) Privacy-safe review summary

About this resource

Knowledge Freshness

This capability pack is grounded in Claude Code sub-agents, MCP, skills, and features overview documentation verified on 2026-06-14. Subagent behavior and MCP inheritance rules can change; prefer live official docs over cached assumptions about tool access.

Retrieval Sources

Source Verification Notes

Verified against official Claude Code sub-agents and MCP documentation and the public Anthropic claude-code repository on 2026-06-14:

  • Claude Code subagents run in isolated context windows for delegated research, exploration, or parallel task slices.
  • MCP tool names and schemas load into parent session context at startup; subagent delegations inherit the broader tool surface available to the session unless restricted by configuration or task design.
  • Claude Code supports MCP tool approval requirements that should be applied before delegating write or execute-class tools to background subagents.
  • Subagents help isolate large reads from the main window but do not automatically isolate MCP side effects on external systems.
  • Official sub-agent documentation describes foreground and background usage patterns that affect how much interactive oversight is practical during tool use.

Scope Note

This is not a replacement for remote MCP server trust review. Use it as a reusable workflow for reviewing MCP scope before subagent delegation in Claude Code.

Core Workflow

  1. Define the delegation boundary: task goal, expected outputs, and forbidden actions.
  2. Inventory parent session MCP servers and classify tools as read, write, execute, or admin.
  3. Map required tools to the delegated task; flag any write/execute tools not strictly needed.
  4. Choose delegation mode: foreground for high-risk tools, background only for read-only research.
  5. Apply MCP scoping mitigations: disable unused servers, enable approval gates, or split task so subagent needs fewer tools.
  6. Review subagent prompt for accidental exfiltration requests such as paste full database dumps into parent chat.
  7. Run a pilot delegation with read-only tools and verify subagent transcript size and sensitive fields.
  8. Document rollback steps if a subagent invokes an unintended MCP write tool.
  9. Produce a privacy-safe scoping summary for the team or incident record.

Capability Scope

  • Parent versus subagent MCP inheritance review.
  • High-risk tool exposure assessment for delegated tasks.
  • Foreground versus background delegation recommendations.
  • MCP scope reduction and approval gate planning.
  • Pilot and rollback procedures.
  • Privacy-safe scoping reporting.

Compatibility

Native

  • Claude Code / Claude: use as an Agent Skill before spawning subagents for research, code exploration, or parallel implementation tasks with MCP enabled.

Manual Adaptation

  • Codex, Cursor, Windsurf, and Generic AGENTS workflows: use the workflow as a deterministic delegation MCP checklist in platform runbooks.

Required Inputs

  • Delegated task description and expected deliverables.
  • Parent session MCP server list and tool classifications.
  • Foreground or background delegation choice and approval policy.
  • Known production systems reachable through MCP tools.

Production Rules

  • Do not background-delegate tasks that require write or execute MCP tools without approval gates.
  • Remove unused MCP servers from the parent session before wide parallel subagent use.
  • Prefer read-only subagent research with summarized outputs over raw tool dumps in parent chat.
  • Treat subagent MCP side effects as production changes requiring the same review as parent tools.
  • Redact customer data from subagent transcripts before sharing externally.
  • Pair this review with remote MCP trust review for third-party servers.
  • Re-scope after MCP config changes mid-session.

Review Matrix

Delegation pattern MCP exposure Recommended mode
Read-only research Low Background OK with summary return
Ticket comment write High Foreground with approval
Database query read Medium Background with redacted summary
Deploy or execute tool Critical Do not delegate; parent only
Many parallel subagents Amplified Remove unused MCP servers first

Output Contract

  1. Parent MCP inventory and subagent inheritance summary.
  2. High-risk tool exposure assessment.
  3. Recommended scope reductions and approval gates.
  4. Foreground versus background delegation decision.
  5. Pilot and rollback procedure.
  6. Privacy-safe review summary.

Duplicate Check

Checked content/skills, content/guides, generated catalog text, and open pull requests for subagent MCP scoping, delegation MCP safety, and inherited tool review workflows. MCP and subagent docs cover features separately, but no skills entry provides a reusable subagent MCP scoping review capability pack with delegation matrix and output contract.

Editorial Disclosure

Submitted as an independent source-backed HeyClaude content entry by kiannidev. It is based on public Claude Code documentation, the public Anthropic claude-code repository, and Google Search Central helpful-content guidance. No paid placement, referral link, affiliate link, or vendor sponsorship is used.

#subagent#mcp#scoping#delegation#capability-pack

Source citations

Add this badge to your README

Show that Subagent MCP Scoping Review Capability Pack Skill is listed on HeyClaude. Paste this Markdown into your README — it renders the badge and links back to this page.

Listed on HeyClaude
[![Listed on HeyClaude](https://heyclau.de/badge/skills/subagent-mcp-scoping-review-capability-pack.svg)](https://heyclau.de/entry/skills/subagent-mcp-scoping-review-capability-pack)

Signals

Loading live community signals…

More like this, weekly

A short, calm digest of reviewed Claude resources. Unsubscribe any time.