Deep reflection skill - read memory and context, decide whether there is something worth sending
中文简介正在准备中。当前显示原文说明。
查看安装方法安装前可以查看交给 AI 代理的指令内容。
Read memory and context to reflect deeply. Design principle: "Reflection first, sending second."
This is an agent-driven skill with no standalone implementation code.
The reflection process is executed by agents/essay-writer.md.
For execution flow diagram, see agents/essay-writer.md → Execution Flow section.
| Source | Agent | Description |
|---|---|---|
/essay | essay-writer.md | Primary invocation |
/essay wait | essay-writer.md | Scheduled (one-time) |
/essay schedule | essay-writer.md | Scheduled (recurring) |
For CLI options, see commands/essay.md → Command Structure section.
First, ingest any replies that have arrived: python main.py replies fetch. This is
plumbing — it moves replies from the inbox onto disk, so what has come back is at hand.
Best-effort: an IMAP, authentication or network failure stops the fetch, not the
reflection — go on to read and write regardless. What is made of them is the writer's;
see What the Plugin Retains.
Read specified files as material for reflection.
Recommended context:
Additional context (AI may read as needed):
Contemplate:
Deliver: There's something worth sharing Don't deliver: Nothing particular to share (this is valid)
| Condition | Mode |
|---|---|
--send flag present | Non-interactive (send email) |
No --send flag | Interactive (display in chat) |
--send flag)Output: Chat display only (no email)
--send flag present)Output: Email (user is not present)
skills/send-email to deliver. Write the essay to a file and send it
with python main.py send --subject-file … --body-file … — the default route, since a body
of more than one paragraph does not fit on a shell argument line. Do not write a throwaway
sending script--in-reply-to '<its Message-ID>' (the ID is in essay_replies.jsonl, and
python main.py replies list prints it beside the subject). The mail then belongs to that
reply's thread instead of arriving as an unrelated letter. Whether an essay is an answer is
the writer's judgment; without the flag it stands on its own, which is the defaultclaude -p directly and never reach the wrapper that wrote essay_wait.log. A --to-self
note, if one is written, lands in the ledger like any other send.The plugin keeps these records on its own account; what becomes of them in a reflection stays with the writer.
essay_ledger.jsonl — one line per sent essay; the bodies stay in sent/ as YYYYMMDD_HHMM.mdessay_replies.jsonl — replies that python main.py replies fetch has ingestedpython main.py send "Subject" "Body" --to-self sends to ESSAY_SENDER_EMAIL and lands in the
same ledger; the recipient field tells a self-addressed note from an essayPaths: skills/send-email/SKILL.md → File Locations.
When sending, write naturally. No forced templates.
Typical elements (not mandatory):
EmailingEssay | GitHub
还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问，或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Development methodology rules (Clean Architecture, TDD flow, 3-Strike rule, decision priority) that govern planning, implementation, and review. Load before designing, staging, implementing, or reviewing code changes.
中文简介正在准备中。当前显示原文说明。
EpisodicRAG 健全性診断と階層推奨
中文简介正在准备中。当前显示原文说明。
EpisodicRAG設定変更（対話的）
中文简介正在准备中。当前显示原文说明。
EpisodicRAG初期セットアップ（対話的）
中文简介正在准备中。当前显示原文说明。
Operational readiness checklist (security, cost, legal, data design, performance, incident response, LLM integration defenses) applied when a change touches deployment, infrastructure, external services, secrets, or user data.
中文简介正在准备中。当前显示原文说明。
Email sending skill (Gmail SMTP + Yagmail)
中文简介正在准备中。当前显示原文说明。