Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
## SecondMind v1.4.0 – "Project Tracker" ### New Features - **Project Tracking**: `/accept` automatically creates tracked projects - `/projects` command to view active/completed projects - `/complete` marks projects as done – permanently excluded from future suggestions - Initiative engine checks project state before suggesting (no duplicates) - Project count in `/status` output ### Fixes - **Initiative dedup overhaul**: Blacklist now covers ALL rejected/dead/completed proposals (was 7 days) - **Hard blacklist safety net**: Fuzzy keyword matching blocks reformulated duplicates even when LLM ignores instructions - **Explicit blacklist section in LLM prompt** with examples and strict rules - `/drop` now archives related knowledge entries to longterm (source dries up, not just the proposal) ### Breaking Changes - None. Migration runs automatically on first start.
What This Skill Does
Ingests OpenClaw conversation sessions and extracts knowledge, emotional context, and events into a three-tier SQLite database. Background jobs run on 30-minute to 6-hour intervals to consolidate data and generate proactive proposals for automations, fixes, and project ideas. Proposals can be accepted, rejected, or deferred via Telegram commands or natural language, and accepted proposals automatically become tracked projects.
Unlike manual note-taking or simple chat history, it runs unsupervised, deduplicates similar insights semantically, and surfaces suggestions without the user needing to query anything.
When to use it
- Searching past sessions for a specific decision or topic
- Reviewing weekly emotional patterns like stress or frustration trends
- Accepting a suggested automation and tracking it as an active project
- Getting reminded of a deferred idea after a configurable cooldown period
- Archiving a long session before resetting the agent context
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: what were my main frustrations this week?
- 1Query the social_context table filtered to the past 7 days
- 2Group results by mood and count occurrences
- 3Format each mood category with its emoji and frequency
- 4Present the ranked list to the user
A ranked mood summary, e.g. 😫 stress x4, 😤 frustration x3, 🤔 curiosity x2
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.