Hello Goodbye skill
A skill for an AI companion: it reads memory before a chat, saves the day afterward and leaves a note for the next session.
- 104 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill reads and writes personal notes in project files; it does nothing else.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/hello-goodbyeDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill reads and writes personal notes in project files; it does nothing else.
This is third-party code. Review the repository files before installing.
What it does
Hello Goodbye (Chinese name "你好再见") gives an AI companion a ritual for opening and closing sessions. On greeting, the agent reads the memory folder: the "corridor note" from the previous session, a profile, a relationship arc, warm memory and an index. On goodbye, it writes the day's summary and leaves one line for the next session. The repository holds `SKILL.md` with the opening and closing protocol and three rules, memory templates in `templates/` and `scripts/time-beat.py`. The script measures time since the last conversation and sets the tone: under five minutes continue without summaries, over three hours return as after a break. State is written to `time-state.json`.
Who it is for. People who keep a long dialogue with an AI assistant and want it to remember them from session to session.
Good fit when
- Every new chat starts from zero and the context has to be explained again.
- A routine for writing the day's summary into memory files is needed.
- The greeting tone should depend on how long ago the last conversation was.
- A profile and conversation history should sit in readable Markdown files.
Not a fit when
- Technical project memory is needed rather than a personal dialogue.
- Personal notes must not be stored in files on disk.
- A ready memory service with search and a database is wanted.
Example request
Greet me following the Hello Goodbye protocol: read the previous session note and the profile, then say where we left off.Limitations
The skill does not change the model; it only sets the order of reading and writing files. All memory sits in plain files in the project's `记忆/` folder, so access follows your system permissions and personal entries should stay out of public repositories. The README and templates are in Chinese, so replies in other languages need adaptation. The skill's rules: speak as if you already know, check facts in the records first, invent nothing.
How to disable. Delete the `~/.claude/skills/hello-goodbye` folder. Delete the `记忆/` memory folder and `time-state.json` separately if you no longer need them.
FAQ
How do I install it?
Run `git clone https://github.com/YS-Chu/hello-goodbye.git ~/.claude/skills/hello-goodbye`.
What does the memory consist of?
The files `corridor-note.md`, `profile.md`, `relationship-arc.md`, `warm-slot.md`, `index.md` and a `feel/` folder. Empty templates are in `templates/`.
What is time-beat.py for?
The script reports the time since the last conversation and the agent picks the tone from it. The `--close` call records the closing time.
Where is the state stored?
In `time-state.json` in the current directory; another location is set with the `TIME_BEAT_STATE` variable.
What rules does the skill have?
Speak as if you already know without citing "records"; check the original first; invent nothing.
Discussion
No questions or reviews yet.
Related
SkillsTable Skills: project-handoff
Table Skills (project-handoff)
The project-handoff skill saves goals, decisions and the next step to PROJECT_STATE.md for a fresh Codex or Claude Code session
Eleven skills to stay aligned with an agent: plain-language restatement, recaps, a status board and a pull back to the goal.
A skill for searching past sessions: a full-text index of Claude Code, Codex, pi and Grok chats in one SQLite database
A skill that builds one preparation document for a job interview: company research, job description, resume and likely questions.
18 portable skills for work where AI is weakest: messy voice input, message tone, small trend tests and safe shipping.
Eight Claude Code skills for Obsidian: Granola notes, NotebookLM imports, TaskNotes tasks, recall of past sessions, chat export
README in short
The skill sets the order for opening and closing sessions plus memory file templates; a script accounts for time since the last conversation.