# Fmind Freelance AI Architect Website

> Read-only access to Fmind's portfolio, articles, and sites.

- name: io.github.fmind/portfolio
- version: 2.5.0
- connection class: R0 (autonomous)
- trust: 97/100
- quality: 83/100 (good)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/io.github.fmind%2Fportfolio
- tools: https://api.protogrid.dev/v1/servers/io.github.fmind%2Fportfolio/tools

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

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "portfolio": {
      "type": "http",
      "url": "https://www.fmind.dev/mcp"
    }
  }
}
```


## Trust

- hygiene: 100
- liveness: 100
- freshness: 100
- provenance: 90
- drivers: +repository +namespace-matches-repo +website +reachable +uptime-100% +updated-5d-ago
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 83/100 (good)

### protocol: 100 (weight 25)

- pass `protocol.modern`: supports 2026-07-28 (server/discover)
- pass `protocol.stateless`: answers without a session
- pass `protocol.transport`: streamable HTTP
- pass `protocol.list_ttl`: tool list is cacheable (ttlMs 3600000)

### 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: 86 (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
- pass `tools.annotations`: 10 of 10 tools declare readOnlyHint or destructiveHint
- fail `tools.directory_hints`: no tool declares all four of readOnlyHint, destructiveHint, idempotentHint and openWorldHint; missing: get_profile (destructiveHint, idempotentHint); list_experience (destructiveHint, idempotentHint); list_certifications (destructiveHint, idempotentHint); …
- pass `tools.token_cost`: about 1,228 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

### stability: 50 (weight 15)

- pass `stability.changes`: 4 tool changes in 30 days
- fail `stability.rug_pull`: 2 tools kept their name but changed most of their description in 30 days; review before trusting them

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

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

- tools: 10, about 1,228 tokens to load them all
- tool set hash: b0211df91b13d9d5c72239a6bca7e43d
- tools last changed: 2026-10-14T11:04:49.240Z
- badge: [![protogrid quality](https://protogrid.dev/badge/io.github.fmind/portfolio.svg)](https://protogrid.dev/servers/io.github.fmind/portfolio)
- Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed.

### Tool changes

- 2026-10-14 `compare_mcp_context_cost` changed (description, inputSchema, outputSchema) **meaning changed**
  - before: Estimate tool-definition tokens, context share and USD costs by filtering, search, CLI and cache assumptions. Call with parameters={} for defaults, bounds and dated sources; edit string-valued URL parameters. Reference tokens and partial-tool estimates are not provider billing measurements. Returns a shareable scenario URL. Invalid inputs fail. See /agents#mcp-context.
  - after: Count the tool-definition tokens that measured MCP servers (GitHub default toolset, Filesystem, Memory, Playwright) add to every model request, then price them as standard input for the managed APIs in the hosting snapshot. Omit servers for all four. Returns tokens per request and month, context share and USD per 1,000 requests and per 30-day month for each model, dated sources, limitations and a shareable page URL. Reference tokens, not provider billing. See /agents#mcp-context.
- 2026-10-13 `get_profile` changed (outputSchema)
- 2026-10-13 `compare_mcp_context_cost` added
  - Estimate tool-definition tokens, context share and USD costs by filtering, search, CLI and cache assumptions. Call with parameters={} for defaults, bounds and dated sources; edit string-valued URL parameters. Reference tokens and partial-tool estimates are not provider billing measurements. Returns a shareable scenario URL. Invalid inputs fail. See /agents#mcp-context.
- 2026-10-12 `compare_llm_hosting` changed (description, outputSchema) **meaning changed**
  - before: Compare one GKE hosting scenario with managed API baselines using the website calculator. Call with parameters={} for defaults and supported model, node, billing, and quantization choices. Then override URL parameters as strings, e.g. requests=1000, throughput=100, tokens=500. Returns assumptions, USD costs, constraints, dated sources, and a shareable URL. Invalid inputs fail; no resources are provisioned. See /agents#calculator.
  - after: Compare a GKE infrastructure budget with managed APIs. Call with parameters={} for defaults and supported models, hardware, and plans. Simple mode accepts preset, model, billing, requests (calls/day), and window (auto, 24, 8, or 1 hour/day); set mode=simple when overriding requests. It sizes replicas under illustrative throughput assumptions. Interactive presets run 24/7; job and pipeline presets schedule running hours inside active windows. Commitments bill idle hours. Advanced mode keeps the fleet fixed and exposes custom token budgets, throughput, and running time. Break-even holds that fleet fixed and checks its capacity. Totals include compute and GKE; operations and additional services are excluded. Returns assumptions, USD costs, constraints, sources, and a shareable URL. Invalid inputs fail; no resources are provisioned. See /agents#calculator.

Full history: https://api.protogrid.dev/v1/servers/io.github.fmind%2Fportfolio/changes

## Remotes

- https://www.fmind.dev/mcp (streamable-http, auth none, reachable true, uptime30d 1)

## Packages

_none_

## Tools (10)

- `compare_llm_hosting`: Compare a GKE infrastructure budget with managed APIs. Call with parameters={} for defaults and supported models, hardware, and plans. Simple mode accepts preset, model, billing, requests (calls/day), and window (auto, 24, 8, or 1 hour/day); set mode=simple when overriding requests. It sizes replicas under illustrative throughput assumptions. Interactive presets run 24/7; job and pipeline presets schedule running hours inside active windows. Commitments bill idle hours. Advanced mode keeps the fleet fixed and exposes custom token budgets, throughput, and running time. Break-even holds that fleet fixed and checks its capacity. Totals include compute and GKE; operations and additional services are excluded. Returns assumptions, USD costs, constraints, sources, and a shareable URL. Invalid inputs fail; no resources are provisioned. See /agents#calculator.
- `compare_mcp_context_cost`: Count the tool-definition tokens that measured MCP servers (GitHub default toolset, Filesystem, Memory, Playwright) add to every model request, then price them as standard input for the managed APIs in the hosting snapshot. Omit servers for all four. Returns tokens per request and month, context share and USD per 1,000 requests and per 30-day month for each model, dated sources, limitations and a shareable page URL. Reference tokens, not provider billing. See /agents#mcp-context.
- `get_article`: Read a public article's complete Markdown, dates, canonical URL, and hosted section links. Use a slug from search_articles; unknown and draft articles are unavailable.
- `get_profile`: Return the core profile: identity, headline, job title, contact, socials, biography, and areas of expertise.
- `get_services`: List the professional services on offer (AI architecture advisory and mentoring) with availability and how to book.
- `list_certifications`: List professional certifications, badges, and course specializations with their issuers.
- `list_experience`: List professional work experience: companies, roles, descriptions, and skill tags.
- `list_projects`: List open-source projects and educational YouTube series.
- `list_publications`: List academic and written publications: the PhD thesis, peer-reviewed papers, and hosted articles.
- `search_articles`: Search the hosted articles by relevance (BM25 over titles, tags, descriptions, and full text) and return the best matches. Use it to answer what Fmind has written about a topic instead of listing every publication.