ha-derived-sensor-authorlisted

Authors one Home Assistant derived or statistical helper sensor as a spec-conformant YAML block from a described intent — bayesian, derivative, filter, min_max, statistics, threshold, trend, history_stats, integration, utility_meter, or group — conforming to the matching spec/ha-automation/<topic>. Picks the right integration, sets the math-bearing parameter (observations/prob_given_*, unit_time, window_size, state_characteristic, lower/upper/hysteresis, min_gradient, method, cycle), types the produced sensor and guards source unavailability, and reports runtime dependencies (recorder retention, integration→utility_meter). Activate on "add a sensor for the rate of change / energy from power / moving average / threshold / trend of…", "make a utility_meter / statistics / bayesian sensor for…". Do not activate for the generic template integration (ha-automation-author), stateful helpers (ha-helper-scaffold), real integration sensors (ha-integration-scaffold), blueprints, or deploying to a live HA instance.
nolte/claude-home-assistant · ★ 1 · AI & Automation · score 68

Install: claude install-skill nolte/claude-home-assistant

# HA Derived Sensor Author Spec: `spec/claude/ha-derived-sensor-author/en.md` (EN canonical) / `spec/claude/ha-derived-sensor-author/de.md` (DE translation). ## Why this is a skill, not an agent - **Human-visible authoring surface** — the user describes the derived value they need and reads back the YAML and the conformance report; a skill keeps this on the visible command surface, like the sibling author/scaffold skills. - **Mid-flow interactivity** — integration selection and the delimitation redirect (rate vs. smoothing vs. integral; threshold vs. trend) are per-run dialogues the user approves before generation. - **Bounded, inline generation** — a single sensor block is small enough to generate inline. - Counter-dimension considered: the draft→validate loop could be an agent, but the integration choice and the runtime-dependency notes belong in the user's working context; skill wins. ## When this skill activates Use this skill to author **one** derived/statistical helper sensor from a described intent: `bayesian`, `derivative`, `filter`, `min_max`, `statistics`, `threshold`, `trend`, `history_stats`, `integration` (Riemann), `utility_meter`, or `group`. ## When NOT to activate - the generic `template:` integration (free-form Jinja sensors) → `ha-automation-author` - a stateful helper (`input_*`, `counter`, `timer`, `schedule`) → `ha-helper-scaffold` - an automation, script, or scene → `ha-automation-author` - a real sensor from your own integration → `ha-integratio