Remind
v1.0.0Auto-learns when and how to bring things back to your human's attention. Adapts timing and style to their preferences.
⭐ 3· 2.4k·13 current·13 all-time
byIván@ivangdavila
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
medium confidencePurpose & Capability
Name and description match the instructions: detect known commitments, decide timing/style, remind, observe reactions, and update preferences. No unexpected binaries, credentials, or installs are requested that would contradict a reminder assistant.
Instruction Scope
SKILL.md, timing.md, and triggers.md stay on topic (detect remindable items, lead times, triggers). However the skill repeatedly says it will 'observe' user reactions and 'auto-learn' without specifying the data sources (conversation history, calendar, email, notifications) or the mechanism for observation. That vagueness means the agent could rely on any conversational context it has access to — reasonable, but worth clarifying before trusting autonomous learning.
Install Mechanism
Instruction-only skill with no install spec or code files. Lowest-risk install profile: nothing is downloaded or written by the skill itself.
Credentials
No environment variables, no credentials, and no required config paths are declared. The requested permissions are proportionate to a reminder assistant. Note: actual operation may require connectors (calendar, email, notification channels) in practice — those are not declared here.
Persistence & Privilege
always is false and model invocation is allowed (normal). The skill says it will 'Store — Update preferences below' but does not specify where preferences are persisted (agent memory, user storage, external service). Ask how and where learned preferences are saved and how to clear them.
Assessment
This skill is instruction-only and appears coherent for a reminder assistant. Before installing, confirm two things with the skill author or your platform: (1) what data sources the skill will 'observe' to learn (conversation history, calendar, email, notifications, etc.), and whether you must grant any connectors; and (2) where learned preferences and reminders are stored and how you can review or delete them. If you are concerned about autonomous learning, run it with model invocation limited or in a restricted test environment until you verify its behavior. If you plan to integrate it with calendars or email, only grant the minimum connector access required.Like a lobster shell, security has layers — review code before you run it.
latestvk97b710yjr9w0856py2nwbevph8108gj
License
MIT-0
Free to use, modify, and redistribute. No attribution required.