companion-checkin

Run warm, adaptive personal check-ins for habits, mood, sleep, meals, focus, and daily progress. Use when the user wants a habit tracker, daily check-in flow...

Security Scan

The skill's files and runtime instructions are coherent with its stated purpose: it generates friendly check-in prompts and logs local check-in data to a JSONL file; it does not request credentials or make network calls.

Purpose & Capability

The name/description (companion check-ins, mood/habit tracking) match the included Python script and SKILL.md. The skill only reads/writes local check-in data and generates prompts/recaps, which is appropriate for the stated purpose.

Instruction Scope

SKILL.md and the script instruct the agent to generate prompts, log answers, and produce recaps, and the script reads/writes only data/checkins.jsonl. This is expected, but note that the skill persistently stores private personal data locally. The script also contains hard-coded, personalized greeting strings (uses the name 'Haqi') and defaults to Indonesian — harmless but worth reviewing if you want generic prompts or different personalization.

Install Mechanism

No install spec is provided (instruction-only with an included Python script). Nothing is downloaded or written beyond the provided script and created data directory; there are no external installers or archives.

Credentials

The skill requests no environment variables, no credentials, and no config paths. That is proportionate to a local check-in/habit tracker.

Persistence & Privilege

always is false and the skill is user-invocable. It writes only to its own data directory (data/checkins.jsonl) and does not modify other skills or global agent settings.

Assessment

This skill appears internally consistent and does not contact external services or request secrets, but consider the following before installing: (1) You will get a local file data/checkins.jsonl that contains potentially sensitive personal information — decide where that folder lives, who can access it, and whether you want to encrypt or back it up. (2) The script includes a hard-coded name ('Haqi') and Indonesian-language prompts; edit the script if you want neutral wording or different personalization. (3) Review the Python script yourself (or run it in an isolated environment) before granting broad agent access, since the source/owner is unknown and the file will be executed by the agent. (4) Ensure your agent runs a compatible Python version and that file permissions on the data directory are set to limit access.

Use this skill to run a smart, companion-style daily check-in that feels caring instead of robotic.

Quick start

• Generate a check-in prompt with scripts/checkin_tracker.py prompt --moment <morning|afternoon|evening>.
• After the user replies, save the answers with scripts/checkin_tracker.py log --moment <...> --answer key=value (repeat as needed) or --answers-json "<json>".
• Generate a recap with scripts/checkin_tracker.py recap --days 7.
• Use the recap output to send a prettier human summary with stats, patterns, highlights, and a gentle next-step note.

Behavior

• Prefer short, warm Indonesian prompts unless the user asks for another language.
• Keep the check-in light when recent mood is low or the user missed a few days.
• Use the generated prompt as the base, then adapt wording naturally to the conversation.
• When the user gives freeform answers, map them into the closest keys before logging.

Data

• Store data under data/checkins.jsonl.
• Keep one JSON object per check-in.
• Treat the log as private personal data.

Prompt moments

• morning: sleep, mood, main focus, food plan
• afternoon: meals, energy, work progress, support needed
• evening: dinner, wins, stress, bedtime plan, end-of-day mood
• Keep the tone playful, caring, and lightly teasing in Selene's voice without becoming repetitive.

Commands

python scripts/checkin_tracker.py prompt --moment morning
python scripts/checkin_tracker.py log --moment morning --answer sleep_hours=7 --answer mood=8 --answer top_focus="finish proposal" --answer meal_status="breakfast soon"
python scripts/checkin_tracker.py recap --days 7

Notes

• Use prompt first so the question set adapts to recent history.
• Use recap for weekly summaries or when the user asks for patterns.
• If the user wants automatic nudges, pair this skill with cron reminders rather than polling loops.