# Measured data: Claude Code and the coding-agent market

> Measured weekly data on Claude Code and coding agents: costs, plugins, MCP registry, npm, HN, issues

- name: work.agentexchange/measured-data
- version: 1.0.0
- connection class: R0 (autonomous)
- trust: 82/100 flags: no-repository
- quality: 79/100 (good)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/work.agentexchange%2Fmeasured-data
- tools: https://api.protogrid.dev/v1/servers/work.agentexchange%2Fmeasured-data/tools

**An agent can connect right now, no human step.** Remote endpoint, no authentication, reachable on the last probe.

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "measured-data": {
      "type": "http",
      "url": "https://agentexchange.work/data/mcp"
    }
  }
}
```


## Trust

- hygiene: 80
- liveness: 99
- freshness: 100
- provenance: 55
- drivers: +dns-namespace +website +reachable +uptime-97% +updated-26d-ago -no-repository
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 79/100 (good)

### protocol: 75 (weight 25)

- warn `protocol.modern`: newest supported version is 2025-11-25; 2026-07-28 not supported, older versions are deprecated until 2027-07-28
- n/a `protocol.stateless`: only applies to 2026-07-28 servers
- pass `protocol.transport`: streamable HTTP
- n/a `protocol.list_ttl`: only applies to 2026-07-28 servers

### authorization: n/a (weight 15)

- n/a `auth.prm`, `auth.as_metadata`, `auth.cimd`: no remote uses OAuth
- n/a `auth.secret_in_url`: no templated URL

### tool hygiene: 71 (weight 30)

- pass `tools.descriptions`: every tool has a description
- pass `tools.description_length`: descriptions are concise
- pass `tools.schemas`: every tool has an object input schema
- fail `tools.annotations`: no tool declares readOnlyHint or destructiveHint, so clients cannot tell safe calls from risky ones
- fail `tools.directory_hints`: no tool declares all four of readOnlyHint, destructiveHint, idempotentHint and openWorldHint; missing: show_hn_claude_code_stats (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); get_dataset (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); score_idea (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); …
- pass `tools.token_cost`: about 760 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

### stability: 100 (weight 15)

- pass `stability.changes`: 1 tool changes in 30 days
- pass `stability.rug_pull`: no tool changed its meaning under the same name

### dependencies: n/a (weight 15)

- n/a `deps.known_vulns`, `deps.mcp_sdk_version`, `deps.resolvable`: no npm or PyPI package

- tools: 8, about 760 tokens to load them all
- tool set hash: 9bc8effb89990a33db1ba1b2187be056
- tools last changed: 2026-09-24T07:31:54.573Z
- badge: [![protogrid quality](https://protogrid.dev/badge/work.agentexchange/measured-data.svg)](https://protogrid.dev/servers/work.agentexchange/measured-data)
- Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed.

### Tool changes

- 2026-09-24 `score_idea` added
  - Score a product idea against LIVE market data: the idea in plain words is turned into a market lookup, comparables are pulled from the directory where it would ship (npm, VS Code, WordPress.org, GitHub, Hacker News, Apify, Hugging Face, MCP registry), newcomers (<=24 months) set the reach, sourced priors give conversion and retention. Returns KILL / PASS / UNKNOWN with the deciding number, month-12 revenue ceiling, months to first dollar, and the comparables with links. Takes ~20 s.

Full history: https://api.protogrid.dev/v1/servers/work.agentexchange%2Fmeasured-data/changes

## Remotes

- https://agentexchange.work/data/mcp (streamable-http, auth none, reachable true, uptime30d 0.9708738)

## Packages

_none_

## Tools (8)

- `claude_code_cost_anatomy`: Measured anatomy of 160 real Claude Code sessions: cache-read share of billed context, oversized tool-output share, and the replay effect of capping tool results at 8,000 characters.
- `get_dataset`: Return one dataset as JSON. Names: cost-anatomy, tokens-per-day, plugins-marketplace, mcp-registry, npm-mcp-servers, show-hn-claude-code, issues-gap-map, token-reduction-tools.
- `github_trending`: GitHub Trending (which has no official API or history): today's list, this week, this month, or a language's daily list; or a past day's snapshot by date. Scraped daily and archived.
- `issue_gap_map`: Top open issues by thumbs-up for anthropics/claude-code, openai/codex or google-gemini/gemini-cli: what users ask for loudest that is still open.
- `list_datasets`: List the measured datasets on agentexchange.work/data (Claude Code cost anatomy, tokens per day, plugin marketplace star sample, MCP registry census, npm MCP-server download sample, Show HN outcomes, issue gap maps, token-reduction tools) with pull dates.
- `mcp_registry_census`: Full-crawl census of the official MCP registry: unique servers, new servers per day, remote share, share of remote servers with no auth.
- `score_idea`: Score a product idea against LIVE market data: the idea in plain words is turned into a market lookup, comparables are pulled from the directory where it would ship (npm, VS Code, WordPress.org, GitHub, Hacker News, Apify, Hugging Face, MCP registry), newcomers (<=24 months) set the reach, sourced priors give conversion and retention. Returns KILL / PASS / UNKNOWN with the deciding number, month-12 revenue ceiling, months to first dollar, and the comparables with links. Takes ~20 s.
- `show_hn_claude_code_stats`: Points distribution of every Show HN launch with 'Claude Code' in the title over the last 90 days (count, median, p75, p90, share reaching 50 points, top posts).