Agent skill

cli-mcp-decision

Decide whether a workflow should use CLI skills, MCP, or a hybrid approach. Use when planning agentic tooling, writing ADRs, or evaluating migrations between CLI and MCP.

View SKILL.md on GitHub Repository
Stars 0
Forks 0

Install this agent skill to your Project

npx add-skill https://github.com/nathanvale/side-quest-marketplace/tree/main/.claude/skills/cli-mcp-decision

SKILL.md

CLI MCP Decision

Use this skill to make repeatable, evidence-based integration decisions for agent tooling.

Quick Start

  1. Capture the workflow context and constraints.
  2. Apply MCP trigger scoring.
  3. Produce a recommendation with rationale, risks, and review cadence.

Inputs To Collect

  • Workflow name and owner
  • Clients that must be supported (Claude-only or multi-client)
  • State requirements (stateless vs session/connection lifecycle)
  • Contract requirements (typed schema, compliance constraints)
  • Reuse scope (single team vs multi-team platform capability)
  • Cost constraints (token budget, latency budget)

Decision Process

Step 1: Start From Default

Assume CLI/skills first for local, stateless, single-client workflows.

Step 2: Evaluate MCP Triggers

Count each trigger that applies:

  • Cross-client interoperability is required
  • Stateful sessions/connections are required
  • Typed contracts/schema validation are required
  • Capability is strategic and reused across teams

Step 3: Apply Crossover Rule

  • If 2+ triggers apply: recommend MCP
  • If <2 triggers apply: recommend CLI/skills
  • If requirements split by workflow phase: recommend Hybrid

Step 4: Validate Non-Functional Constraints

  • Compare expected token overhead and latency
  • Document reliability/fallback behavior
  • Document security and ownership boundaries

Step 5: Output Decision Record

Return a concise decision block:

  • Recommendation (CLI/skills, MCP, or Hybrid)
  • Trigger score (N/4)
  • Top 3 reasons
  • Key risks and mitigations
  • Reassessment date (default: 90 days)

Output Template

md
Decision: <CLI/skills|MCP|Hybrid>
Trigger score: <N>/4
Reasons:
1. ...
2. ...
3. ...
Risks:
- ...
Mitigations:
- ...
Reassess on: <YYYY-MM-DD>

Done When

  • Recommendation is explicit and trigger-scored
  • Tradeoffs are documented (token, latency, reliability, portability)
  • Reassessment date is set

Recommended Agent Skills

Expand your agent's capabilities with these related and highly-rated skills.

nathanvale/side-quest-marketplace

agent-cli

Patterns for building CLIs that serve both humans and AI agents. Covers tri-modal output (JSON/JSONL/human), typed exit codes, structured error contracts, zero-dependency arg parsing, topic-based help, service discovery, field projection, and auto agent-mode detection. Reference implementation: @side-quest/observability CLI. Use when: building a CLI, adding a command, CLI output format, agent-friendly CLI, machine-readable output, JSON CLI contract, exit codes, --json flag, NDJSON streaming, how should my CLI work with agents, CLI design patterns.

0 0
Explore
nathanvale/side-quest-marketplace

explain-codebase

Generate an engaging, deep-dive explanation document (FOR_[NAME].md) that explains a codebase or domain in plain language with analogies, lessons learned, and practical wisdom.

0 0
Explore
nathanvale/side-quest-marketplace

vscode-bookmarks

Create and manage VS Code Bookmarks extension labeled bookmarks in .vscode/bookmarks.json. Use when the user wants to add navigation bookmarks to a codebase, trace data flows through code, mark important code locations, or set up labeled waypoints for code walkthroughs. Also use when user says "bookmark", "label this code", "mark this flow", or "add waypoints".

0 0
Explore
nathanvale/side-quest-marketplace

manage-plugins

Add or remove inline plugins from the ccdev function in dotfiles .zshrc

0 0
Explore
nathanvale/side-quest-marketplace

observability

Start, stop, or check status of the Side Quest observability event server. Use when asked to start the server, stop the server, check server status, or manage the observability dashboard lifecycle.

0 0
Explore
nathanvale/side-quest-marketplace

node-cert

Fix Node.js certificate trust issues behind corporate SSL-inspecting proxies. Use when discussing: VPN, corporate VPN, corporate proxy, SSL inspection, MITM proxy, TLS interception, NODE_EXTRA_CA_CERTS, self-signed certificate error, UNABLE_TO_VERIFY_LEAF_SIGNATURE, certificate chain, proxy CA, corporate network, Claude Code on VPN, Gemini CLI on VPN, npm behind proxy, Node.js certificate error, CAFile.pem, proxy-on, proxy-off, HTTPS_PROXY, HTTP_PROXY, openssl s_client, certificate extraction, corporate root CA, forward trust CA, trust store gap.

0 0
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results