iRobot Roomba and Braava

Before you start: read this whole install prompt, tell me in a few lines what this skill does and what installing it involves, and wait for my go-ahead before you do anything else. Prerequisites — Muse Home Link and a compatible physical device are required to USE this skill, not to install it: installing the skill file itself needs no account, no key and no device, and this install does NOT set up Home Link, pair anything or ask for any credential. But the skill is instructions only. To use it I must read and follow ~/docs/devices/home_link.md in my Muse environment, have the compatible physical device named below already set up on my local network, and have the robot's existing BLID and local password in approved secret storage, a commissioned robot, and my explicit permission for a cleaning, pause or dock action with the area safe for movement. Installing the file configures none of that. Install "iRobot Roomba and Braava" for me. It works with a compatible Wi-Fi Roomba or Braava whose exact model and firmware support the documented local protocol — the presence of TCP 8883 alone is not sufficient to identify an iRobot or prove local control — using the current local endpoint. With a dorita980 local client (or equivalent MQTT/TLS implementation) routed through Home Link, retaining device-specific TLS handling, it instantiates the local interface with the current endpoint, the robot's existing BLID and local password and the appropriate protocol version. It authenticates, subscribes to the supported state feed, and reads mission/cleaning state, battery, docking and reported errors before any command. It sends only the model's documented requested cleaning, pause, resume or return-to-dock command, checking whether the current mission state requires a particular sequence rather than issuing generic commands blindly. It monitors fresh mission state and releases the session when finished, avoiding competing sessions. Verification means the reported mission actually changed, not merely that MQTT accepted a message; returning to base differs from docked/charging, paused from completed. Limits: no cloud credential retrieval, no UDP discovery from the agent, no new provisioning, no universal model support; room maps, targeted jobs, mop/water controls and station functions are outside scope unless documented for the exact model and already configured; physical, battery, bin and safety conditions are never bypassed. Repository: https://github.com/facebookincubator/muse-gadget-sdk (the skill is the whole skills/gadget-irobot-roomba-braava/ folder at the pinned commit: SKILL.md only, 1 file) Pinned version: install the code at commit e6fe5174092702ffddae6f90759a38d55a4ebae1 exactly. Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Clone https://github.com/facebookincubator/muse-gadget-sdk into a temporary folder, check out commit e6fe5174092702ffddae6f90759a38d55a4ebae1, and copy the skills/gadget-irobot-roomba-braava/ folder from that checkout. Then compare every file in the folder with the fingerprints recorded for this listing on 2026-10-22: [FINGERPRINTS] - SKILL.md: SHA-256 c6c2bfd5ce81810ea448fc954efc4507a0a02f892754efffe27e436140a51928 If a file does not match, it changed since the listing was prepared: do not install it yet, tell me what you found and let me decide. 2. Read SKILL.md in full, and tell me in plain words: what the skill does when used, what it connects to, whether it reads or sends any personal information or credentials, and anything that seems unusual. Stop on any red flag and tell me before going further. Confirm these points from the files: - The skill folder holds 1 file only: SKILL.md (2,558 bytes). There is no script, no code and no configuration in it; at use time it works through a compatible dorita980 local client or equivalent MQTT/TLS implementation routed through Muse Home Link against my own robot, on local TCP 8883. It has no dependency on any other gadget skill in this lot. - Nothing is sent anywhere during installation itself; the only download at install time is the skill folder from GitHub at the pinned commit. No robot is contacted, no credential is asked for and no cleaning action is started during installation. - At use time, exchanges go only to my local robot: the client's local interface is instantiated with the current endpoint and the robot's existing BLID and local password from approved secret storage. The skill expressly forbids invoking cloud login or credential-discovery helpers: no cloud credential retrieval ever happens, and no credential leaves the local network for any third party. - The skill does no UDP discovery from the agent and no new provisioning, and it does not bypass physical, battery, bin or safety conditions; room maps, targeted jobs, mop/water controls and station functions stay outside scope unless documented for the exact model and already configured. - Do not pair any device, do not ask for or set any credential, do not run any control, print, scan or playback action, and do not use the skill in any way during this install: Home Link setup, pairing and use are my decisions, later, when I actually ask for them. 3. When I say "yes, install", install it as a skill: copy the whole gadget-irobot-roomba-braava folder into the agent's skills directory, in a folder named exactly "gadget-irobot-roomba-braava", keeping its 1 file (SKILL.md at its root). The exact folder name matters: relative links between sibling gadget skills depend on these exact names. 4. Verify with no network calls: frontmatter valid (name: gadget-irobot-roomba-braava), the 1 file in place with matching fingerprint. Do not run anything and do not connect to any device to verify. 5. Report what was installed, where, and what I still need to do myself (for example: read ~/docs/devices/home_link.md in my Muse environment and follow its Home Link rules; have a commissioned, compatible Wi-Fi Roomba or Braava on my local network; have its existing BLID and local password in approved secret storage; and give explicit permission for each cleaning, pause or dock action with the area safe for movement — no cloud credential will ever be retrieved; using the skill is a separate step that only starts when I ask for it, and the use-time exchanges disclosed above happen only then). GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. If anything looks off, stop and ask me.