Joanna Idea Scout - 2026-09-21 #1956

Joanna Idea Scout

• Date: 2026-09-21
• Generated: 2026-09-21T07:24:25.521Z
• Digest items: 11
• Sources: GitHub, HA Community

This daily issue keeps all Idea Scout findings in one place for review, regardless of source adapter. These are inspiration items, not bugfixes or support threads.

Ideas

1. [GitHub] CodeWithCJ/SparkyFitness - [Feature]: Add configurable water drinking reminders for mobile app

• Source: [Feature]: Add configurable water drinking reminders for mobile app CodeWithCJ/SparkyFitness#2473
• Strength: strong
• Area: packages
• Summary: Configurable water-intake reminders could extend the existing health-coaching workflow with user-controlled, low-noise hydration prompts.
• Suggested practical implementation: Add a health-coaching package with an input_datetime helper for the last water intake, a waking-hours/quiet-hours helper, a dashboard quick action or Joanna command to record intake, and an automation that sends a Telegram reminder when no intake has been recorded for two hours.
• Potential added value: Provides a concrete hydration prompt with configurable timing while keeping reminders quiet outside selected hours and making intake tracking accessible from the existing health workflow.

2. [GitHub] CodeWithCJ/SparkyFitness - [Feature]: Allow ingredient quantity adjustments before logging a saved meal

• Source: [Feature]: Allow ingredient quantity adjustments before logging a saved meal CodeWithCJ/SparkyFitness#2543
• Strength: strong
• Area: general
• Summary: Allow saved meal ingredients to be adjusted for a single health-log entry without changing the reusable meal template.
• Suggested practical implementation: Extend the health meal-logging workflow with an optional per-entry ingredient override step that accepts changed quantities, recalculates the entry's nutrition totals, and stores the adjusted log separately from the saved meal template.
• Potential added value: Reduces repeated editing after logging, improves nutrition accuracy for variable portions, and preserves reusable meal templates as stable defaults.

3. [GitHub] CodeWithCJ/SparkyFitness - [Feature]: Let the AI Coach update, delete, and edit workout presets

• Source: [Feature]: Let the AI Coach update, delete, and edit workout presets CodeWithCJ/SparkyFitness#2500
• Strength: strong
• Area: general
• Summary: Enable Joanna's health-coaching workflow to manage reusable workout presets with structured exercises, sets, reps, superset groups, numeric identifiers, and confirmation before deletion.
• Suggested practical implementation: Extend the shared health-coaching capability with workout-preset operations that list presets by name or numeric ID, create and update exercise structures including sets, reps, and superset groups, log a selected preset, and require explicit confirmation before deletion.
• Potential added value: Could make recurring workouts easier to maintain and log from Telegram while reducing repeated manual entry and preventing accidental preset deletion.

4. [GitHub] CodeWithCJ/SparkyFitness - [Feature]: Migraine Tracking

• Source: [Feature]: Migraine Tracking CodeWithCJ/SparkyFitness#1882
• Strength: strong
• Area: general
• Summary: Add structured migraine tracking with symptom, severity, medication, trigger, and contextual health data capture plus recurring pattern summaries.
• Suggested practical implementation: Add a Joanna migraine check-in workflow that records onset time, severity, symptoms, aura or nausea, medications, suspected triggers, and notes into a linked OneNote health log, then generates a weekly summary combining each event with the prior day's available hydration, sleep, food, and activity entries.
• Potential added value: This could replace fragmented notes with consistently structured records and make recurring trigger patterns easier to identify without requiring manual correlation.

5. [GitHub] homeassistant-ai/ha-mcp - [FEATURE] Surface pending tool-approval requests outside the settings panel

• Source: [FEATURE] Surface pending tool-approval requests outside the settings panel homeassistant-ai/ha-mcp#2502
• Strength: strong
• Area: automation
• Summary: Expose HA-MCP tool-approval requests as a Home Assistant event so unattended gated actions can trigger local notifications instead of expiring silently.
• Suggested practical implementation: Add a Home Assistant package automation that listens for the proposed ha_mcp_approval_requested event and sends an urgent Joanna/Telegram notification containing the tool name, affected arguments, approval token, and expiry time, while also creating a persistent Home Assistant notification for in-dashboard visibility.
• Potential added value: Makes unattended approval requests visible before their TTL expires, reducing silent denials and improving response time for legitimate security-sensitive actions without notifying on ordinary HA service calls.

6. [GitHub] homeassistant-ai/ha-mcp - feat(ci): maintainer /astra and /sol issue-to-PR workflow

• Source: feat(ci): maintainer slash agent for issues and PRs homeassistant-ai/ha-mcp#2467
• Strength: strong
• Area: general
• Summary: A maintainer-only slash-command workflow can turn authorized GitHub issues into durable, reviewable Codex work sessions, answering directly when no repository change is needed and creating a draft PR when implementation is required. Its separation of read-only execution, checkpointed state, and narrowly scoped publication is relevant to BearClaw maintenance automation.
• Suggested practical implementation: Add a maintainer-only GitHub Actions command for codex_appliance maintenance issues that records a durable checkpoint, runs the coding worker with read-only repository credentials, and publishes validated changes through a separate narrowly scoped publisher into a draft PR; keep merging and reviewer requests manual.
• Potential added value: Could reduce repetitive operator coordination while preserving an auditable checkpoint, safer credential boundaries, resumable review corrections, and explicit human control over final repository integration.

7. [GitHub] jdeath/CheckRoyalCaribbeanPrice - Add optional dining and entertainment reservation-release alerts

• Source: Add optional dining and entertainment reservation-release alerts jdeath/CheckRoyalCaribbeanPrice#138
• Strength: strong
• Area: general
• Summary: Optional release alerts for cruise dining and entertainment reservations could extend the existing Cruise Tracker workflow beyond price monitoring.
• Suggested practical implementation: Add a Cruise Tracker scheduled availability check for configured sailings that detects newly released dining or entertainment slots, suppresses duplicate alerts, and sends a Telegram notification containing the sailing, offering, date/time, and booking link.
• Potential added value: This could reduce manual checking and improve the chance of catching limited reservation inventory while keeping alerts focused on actual release events.

8. [GitHub] RoggeOhta/awesome-codex-cli - Add loopguard — session transcript watchdog (loop/oscillation/stall alerts)

• Source: Add loopguard — session transcript watchdog (loop/oscillation/stall alerts) RoggeOhta/awesome-codex-cli#298
• Strength: strong
• Area: general
• Summary: Add a watchdog for long-running Codex/BearClaw sessions that detects repeated action patterns or stalled progress and notifies Joanna through Telegram.
• Suggested practical implementation: Add a BearClaw watchdog workflow that reads authorized Codex session activity, flags configurable repeated tool/action signatures or inactivity thresholds, and updates the existing temporary Telegram status message or sends an operator alert without changing human-lane routing.
• Potential added value: Could reduce unattended session stalls and make loop or oscillation conditions visible sooner, improving operational response while reusing the existing session and Telegram delivery paths.

9. [GitHub] CodeWithCJ/SparkyFitness - v1.7.2

• Source: https://github.com/CodeWithCJ/SparkyFitness/releases/tag/v1.7.2
• Strength: recap
• Area: automation
• Summary: SparkyFitness v1.7.2 adds configurable hydration reminders, which is a useful pattern for making health coaching prompts adjustable instead of hard-coded.
• Suggested practical implementation: Add a Home Assistant automation that reads a configurable hydration-target helper and reminder interval, then sends Telegram-first Joanna reminders only while the daily target remains incomplete; reset the tracking state at the start of each day.
• Potential added value: This could reduce missed hydration prompts while allowing reminder frequency and daily targets to be tuned without editing automation logic.

10. [GitHub] sdras/awesome-actions - Add AICostFence to Static Analysis

• Source: Add AICostFence to Static Analysis sdras/awesome-actions#959
• Strength: recap
• Area: general
• Summary: AICostFence estimates AI SDK spend from repository traffic assumptions and flags unbounded output or agent tool loops before merge.
• Suggested practical implementation: Add a pull-request workflow in codex_appliance that runs AICostFence against agent-related changes, using repository-owned traffic assumptions to flag unbounded output or tool-loop cost before merge.
• Potential added value: Could reduce unexpected AI usage from runaway agent loops and make cost risks visible during review, with limited maintenance overhead if the action supports the repository's execution model.

11. [HA Community] HA Community Projects - Jev: a decision model for Home Assistant. Sensors, automation actions, and an Assist agent

• Source: https://community.home-assistant.io/t/jev-a-decision-model-for-home-assistant-sensors-automation-actions-and-an-assist-agent/1025669
• Strength: recap
• Area: automation
• Summary: Jev provides typed, confidence-aware decision sensors and automation actions that can replace deeply nested templates when an HA workflow needs classification, scoring, or an explicit uncertain outcome.
• Suggested practical implementation: Add a Home Assistant package automation that passes the existing Frigate/Wyze event context to a Jev decision action, sends high-confidence actionable events through the established notification workflow, records low-confidence results in the logbook or a low-priority channel, and leaves ambiguous cases without automatic escalation.
• Potential added value: Could reduce noisy camera interruptions and make uncertain classifications visible instead of silently dropping them, while centralizing decision logic that would otherwise become nested templates.