skill
personal-agent-growth
A portable skill for task reviews, evidence, corrections, and shared-file handoffs between agent sessions. / 用任务复盘、证据、更正与共享文件衔接智能体会话。
Set up
Skills are installed by placing their folder where your agent loads skills.
The download comes from github.com/KarmenLoop/personal-agent-growth itself, not from our server, so what you get is exactly what its author published.
Install instructions come from the README
README
synced Oct 23, 2026Personal Agent Growth
English | 简体中文
Give the next work session something better than a chat recap. Keep task state, evidence, corrections, and one useful next action in a shared private folder that the next host can actually read.
Download v0.3.3 · Guide · Issues
What it does
A portable instruction skill and offline package checker for task review and handoff. The skill and its four working references are currently Chinese; the public setup guides are available in both languages.
Workflow illustration for v0.3.3; not an application screenshot or runtime acceptance.
Start here
Download the skill ZIP and manifest.json. Extract into a fresh canonical folder, then add that folder once through your host’s supported skill discovery path. Keep the skill folder separate from private task records.
Choose one private task directory that both hosts resolve to the same physical location. Record the active writer and state_revision before a handoff. Python 3.11+ is needed only for the optional offline checker, not for reading Markdown.
A first useful task
Explicitly load SKILL.md, review a synthetic task artifact, save the actual check result, correct one mistaken assumption, and leave a concrete next review action. Confirm those files exist before changing hosts.
Have the receiving host read the current revision and use the earlier artifact in a new check. Scheduling needs an actual supported backend and separate task authorization; installation creates no timer or cross-device sync.
Questions
Does this retrain a model? No. It organizes reusable task records and review methods. Model weights and provider accounts are unchanged.
Documentation and license
Download, setup, practice, and updates · LICENSE
AGPL-3.0-only. The original license terms govern use and redistribution; retain the included legal files. KarmenLoop.