@vk0dev
基本情報
Local-first Claude Code cost analyzer. Parses JSONL session logs to surface per-tool spend, daily trends, and optimization hints.
以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。
{
"mcpServers": {
"agent-cost": {
"command": "npx",
"args": [
"-y",
"@vk0/agent-cost-mcp"
]
}
}
}Parse a single Claude Code session and return token totals, turn count, cache usage, and estimated USD cost so an agent can anchor every later Cost Guard question in one concrete run summary.
Aggregate tool invocations across one session or a filtered project log directory, reporting per-tool call counts and context-share percentages so you can see which tool patterns are actually driving spend.
Roll session logs into a day-by-day local cost trend, with per-day sessions, tokens, and estimated spend so anomalies and rising burn patterns are visible before they feel like guesswork.
Return a parent-plus-subagent session tree for one local Claude Code run, with cost summed per branch, so you can see which branch or delegated path actually consumed the budget.
Rank tools by a bounded ROI heuristic using cost share, linked results, and context share, so repeated calls with weak payoff surface quickly as the classic low-efficiency or runaway-loop signature, while productive same-tool refinement is less likely to be flagged too aggressively after `2.3.1`.
Generate lightweight optimization suggestions from a parsed session log, including cache-read ratios, abandoned tool calls, and heaviest turns, when you want the next fix to be more concrete than a raw metric table.
Flag unusually high or low daily cost spikes against the recent local baseline so sudden burn jumps, suspicious drops, and unstable usage patterns stand out without a separate monitoring stack.
Project a bounded local cost forecast from recent daily trend data so an operator can ask where spend is heading next, not just where it already went; `forecast_confidence` is a quartile-based local heuristic, not certainty, and it degrades gracefully when history is still short.
Set daily or per-session budget caps with tiered alert thresholds, so the next cost-query tool can return a machine-readable warning before an agent quietly runs past a soft or hard spending boundary.
Local Cost Guard and runtime guardrails for AI agents in Claude Code. It reads your local JSONL session logs to show where spend is coming from, forecast where it is heading, attribute cost by provider/model/tool, roll up subagent trees, and let operators push signed monitor-webhook alerts—all without a hosted control plane, cloud, or API key.
Install by adding an npx-based MCP server to your client configuration (Claude Desktop, Claude Code CLI, Cursor, or Cline). No global install required. Once configured, ask the agent “What tools does agent-cost expose?” to see the eleven available tools. Operates on existing session logs at ~/.claude/projects/.
get_session_cost, get_tool_usage, get_cost_trend, get_subagent_treeget_tool_roi, suggest_optimizations, detect_cost_anomaliesget_cost_forecast, estimate_run_costconfigure_budget, set_monitor_webhookget_subagent_treeNative `/cost` gives quick current-session totals and statusline numbers. @vk0/agent Cost Mcp provides per-tool cost share, ROI ranking, anomaly detection, cost forecasts, and agent-readable budget guardrails—all machine-readable from local JSONL logs.
Use `npx -y @vk0/agent-cost-mcp` as the command in your MCP client’s server configuration (e.g., `claude_desktop_config.json`, `.mcp.json`, or `.cursor/mcp.json`). No global installation needed.
The tool reads session logs stored locally at `~/.claude/projects/<project>/`. No data leaves your machine.
No. It is for developers and operators who want Cost Guard–style answers from local Claude Code logs, not for chargeback reporting, company-wide spend accounting, or live runtime introspection.
The server communicates over stdio MCP transport. No authentication or API keys are required.
TruVerifAI's Panel Review is the guardian agent for AI's highest-stakes coding decisions: four frontier models argue over an AI coding agent's riskiest designs, diffs, and commits before they execute, with local review g
HeyLead: AI agent for LinkedIn outreach. Finds the right people, writes in your voice, follows up and handles replies, from Claude Code, Cursor, any MCP client or a web dashboard.
Operate a real LinkedIn account from your AI agent. 52 tools for the inbox, invitations, network, profiles, Sales Navigator search in plain language, posts and engagement; six playbooks served as MCP prompts that preview
NetFoundry LLM Gateway is an OpenAI-compatible API proxy that routes requests across multiple LLM providers using zero-trust networking. Zero trust LLM gateway. OpenAI-compatible proxy with semantic routing and load bala
The AI marketing canvas for Claude and MCP agents. Your agent drafts campaigns as cards on a shared board, briefed by your brand rules, approved by you.
コメント