# hackshop

> Plan and build a physical body for your AI agent: boards, parts list, assembly and pairing.

- name: dev.hackshop/hackshop-mcp
- version: 0.0.6
- connection class: R0 (autonomous)
- trust: 83/100 flags: no-repository
- quality: 73/100 (needs work)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/dev.hackshop%2Fhackshop-mcp
- tools: https://api.protogrid.dev/v1/servers/dev.hackshop%2Fhackshop-mcp/tools

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

## Connection (mcpServers)

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


## Trust

- hygiene: 80
- liveness: 100
- freshness: 100
- provenance: 55
- drivers: +dns-namespace +website +reachable +uptime-100% +updated-16d-ago -no-repository
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 73/100 (needs work)

### 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: assess_hackability (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); get_firmware_playbook (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); verify_firmware_artifact (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); …
- pass `tools.token_cost`: about 2,513 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

### stability: 75 (weight 15)

- warn `stability.changes`: 13 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`: deps.dev does not know hackshop-mcp@0.0.6

- tools: 10, about 2,513 tokens to load them all
- tool set hash: 51d8dc3d012a88c1a9fd23d966f463bd
- tools last changed: 2026-10-14T23:11:16.483Z
- badge: [![protogrid quality](https://protogrid.dev/badge/dev.hackshop/hackshop-mcp.svg)](https://protogrid.dev/servers/dev.hackshop/hackshop-mcp)
- 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 `find_reverse_engineering_examples` added
  - Search Hackshop's curated hardware reverse-engineering examples and research methods. With no query, returns the newest dated examples first. Evidence labels distinguish reproduced work, documented implementations, author demos and analysis methods; they do not imply compatibility with similar-looking hardware.
- 2026-10-08 `prepare_firmware_job` added
  - Prepare a non-executing firmware manifest after owner authorization and an exact compatibility match. Returns backup, recovery, sources, validation and stop points; never returns a flash command.
- 2026-10-08 `verify_firmware_artifact` added
  - Compare a caller-supplied filename, source URL and SHA-256 with Hackshop's curated metadata. Hackshop does not fetch, upload or redistribute the artifact; absent hashes return unknown.
- 2026-10-08 `get_firmware_playbook` added
  - Return a curated firmware playbook with exact targets, prerequisites, backups, recovery, independent risk axes, sources, validation checks and the human-only destructive-action boundary.
- 2026-10-08 `check_firmware_compatibility` added
  - Fail-closed compatibility check for one firmware playbook. Required model, codename, revision, SoC or firmware identifiers must exactly match a curated target; missing facts return unknown and mismatches return unsupported.
- 2026-10-08 `find_firmware_playbooks` added
  - Find curated owner-authorized firmware, local-protocol and cloud-replacement playbooks. Results prefer the least-invasive path and are not compatibility approval; call check_firmware_compatibility before preparation.
- 2026-10-08 `assess_hackability` changed (description)
  - before: Look up a device by id, exact name or substring and return hackability, brick risk, firmware links, community size, build page and agent platform support. Input: `device_name` 1-200 chars.
  - after: Look up a device by id, exact name or substring and return hackability, brick risk, firmware links, curated firmware playbooks, community size, build page and agent platform support. Input: `device_name` 1-200 chars.
- 2026-10-06 `get_build_plan` changed (description)
  - before: Return the complete deterministic build plan for a catalog device id: parts, shopping list with buy_options and store URLs, steps, machine-readable assembly, commands, links, caveats and an agent-ready brief. Input: `device_id` 1-200 chars.
  - after: Return the complete deterministic build plan for a catalog device id: difficulty, board-specific flash warnings, parts, shopping list with buy_options, prices checked date and purchase policy, steps, machine-readable assembly with verify checks, commands, links, caveats and an agent-ready brief. Input: `device_id` 1-200 chars.
- 2026-10-06 `plan_gadget` changed (description, inputSchema)
  - before: Deterministically plan a physical body for an AI agent using Meta Muse boards. Inputs: `idea` 3-2000 chars; `platform` one of muse-esp32, muse-linux, any (default any); `budget_usd` positive number up to 100000; `owned_device_ids` up to 50 catalog ids; `needs` values: voice, screen, images, touch, camera, air-sensors, e-ink, battery, round, home-tunnel, big-screen, compute, linux; `size` one of pocket, desk, wall, hidden, any (default any); `limit` 1-5 (default 3). Returns fit, notes, warnings, intake questions, picks, terms and next steps.
  - after: Deterministically plan a physical body for an AI agent using Meta Muse boards. Ask the intake questions first (intake_gadget, or the `questions` this returns until `intake.complete` is true), then pass the human's choices as `answers`. Inputs: `idea` 3-2000 chars (optional when `answers` or `needs` are given); `answers` {size, interaction, sensing, budget} using the intake option values; `platform` one of muse-esp32, muse-linux, any (default any); `budget_usd` number greater than 0, up to 100000; `owned_device_ids` up to 50 catalog ids; `needs` values: voice, screen, images, touch, camera, air-sensors, e-ink, battery, round, home-tunnel, big-screen, compute, linux; `size` one of pocket, desk, wall, hidden, any (default any); `limit` 1-5 (default 3). Boards that meet the hard needs and the budget rank first, cheapest first. Returns fit, notes, warnings, intake status, questions, picks with difficulty, terms and next steps.
- 2026-10-06 `intake_gadget` added
  - Return the 4 intake questions (where it lives, how you interact, room sensing, budget), why each matters and how each answer maps to plan_gadget. Ask the human these, then call plan_gadget with `answers`. No input.

Full history: https://api.protogrid.dev/v1/servers/dev.hackshop%2Fhackshop-mcp/changes

## Dependencies

- npm hackshop-mcp@0.0.6: unknown

Dependency graphs from deps.dev; advisories from OSV.dev, including the GitHub Advisory Database and the PyPI Advisory Database; all CC BY 4.0.

## Remotes

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

## Packages

- npm hackshop-mcp@0.0.6 secrets: ANTHROPIC_API_KEY

## Tools (10)

- `assess_hackability`: Look up a device by id, exact name or substring and return hackability, brick risk, firmware links, curated firmware playbooks, community size, build page and agent platform support. Input: `device_name` 1-200 chars.
- `check_firmware_compatibility`: Fail-closed compatibility check for one firmware playbook. Required model, codename, revision, SoC or firmware identifiers must exactly match a curated target; missing facts return unknown and mismatches return unsupported.
- `find_firmware_playbooks`: Find curated owner-authorized firmware, local-protocol and cloud-replacement playbooks. Results prefer the least-invasive path and are not compatibility approval; call check_firmware_compatibility before preparation.
- `find_reverse_engineering_examples`: Search Hackshop's curated hardware reverse-engineering examples and research methods. With no query, returns the newest dated examples first. Evidence labels distinguish reproduced work, documented implementations, author demos and analysis methods; they do not imply compatibility with similar-looking hardware.
- `get_build_plan`: Return the complete deterministic build plan for a catalog device id: difficulty, board-specific flash warnings, parts, shopping list with buy_options, prices checked date and purchase policy, steps, machine-readable assembly with verify checks, commands, links, caveats and an agent-ready brief. Input: `device_id` 1-200 chars.
- `get_firmware_playbook`: Return a curated firmware playbook with exact targets, prerequisites, backups, recovery, independent risk axes, sources, validation checks and the human-only destructive-action boundary.
- `intake_gadget`: Return the 4 intake questions (where it lives, how you interact, room sensing, budget), why each matters and how each answer maps to plan_gadget. Ask the human these, then call plan_gadget with `answers`. No input.
- `plan_gadget`: Deterministically plan a physical body for an AI agent using Meta Muse boards. Ask the intake questions first (intake_gadget, or the `questions` this returns until `intake.complete` is true), then pass the human's choices as `answers`. Inputs: `idea` 3-2000 chars (optional when `answers` or `needs` are given); `answers` {size, interaction, sensing, budget} using the intake option values; `platform` one of muse-esp32, muse-linux, any (default any); `budget_usd` number greater than 0, up to 100000; `owned_device_ids` up to 50 catalog ids; `needs` values: voice, screen, images, touch, camera, air-sensors, e-ink, battery, round, home-tunnel, big-screen, compute, linux; `size` one of pocket, desk, wall, hidden, any (default any); `limit` 1-5 (default 3). Boards that meet the hard needs and the budget rank first, cheapest first. Returns fit, notes, warnings, intake status, questions, picks with difficulty, terms and next steps.
- `prepare_firmware_job`: Prepare a non-executing firmware manifest after owner authorization and an exact compatibility match. Returns backup, recovery, sources, validation and stop points; never returns a flash command.
- `verify_firmware_artifact`: Compare a caller-supplied filename, source URL and SHA-256 with Hackshop's curated metadata. Hackshop does not fetch, upload or redistribute the artifact; absent hashes return unknown.