Total Recall
Give your coding agent persistent, automatic session memory without running a vector database or reminding it to save notes.
Overview
total-recall is a journey-wide agent skill that automatically compresses and consolidates conversation memory with an LLM observer—usable whenever a solo builder needs continuity across agent sessions without a vector da
Install
npx skills add https://github.com/monasprox/openclaw-memory --skill SKILL.mdWhat is this skill?
- LLM observer compresses conversations into prioritised notes without manual save commands
- No database, no vectors—file-oriented memory with consolidation when notes grow
- Five layers of redundancy with recovery for missed captures
- Aimed at ~$0.00/month cost when using free-tier models
- Runs maintenance-free compared to skills that require you to remember to remember
- Five layers of redundancy for memory recovery
- No database and no vectors required
- ~$0.00/month when using free-tier models
Adoption & trust: 1 GitHub stars.
What problem does it solve?
Your agent forgets decisions and context between sessions unless you manually save notes or pay for a heavy memory stack.
Who is it for?
Solo builders running long agent sessions who want zero-ops file memory and are fine with observer-based compression on free-tier models.
Skip if: Teams that need enterprise vector search, shared multi-user memory graphs, or strict offline-only air-gapped retention without any LLM observer calls.
When should I use this skill?
You want the agent to remember prior conversations automatically without manual saves, vector infrastructure, or you prompting it to store context.
What do I get? / Deliverables
Prioritised, consolidated notes persist with redundant recovery so later sessions pick up thread without you issuing save rituals.
- Prioritised compressed memory notes on disk
- Consolidated memory after growth thresholds
- Recovered notes when observer misses a turn
Recommended Skills
Journey fit
Useful at every journey phase - explore requirements and options before committing to a direction.
Where it fits
Recall last session’s scope cuts and open questions before you approve a prototype plan.
Keep review preferences and recurring defect patterns across PR threads without re-explaining them.
Retain messaging variants you already rejected so the agent does not resurrect dead copy.
Continue an incident or bug narrative across days with consolidated operator notes.
Remember frequent user complaints and your drafted replies for faster support drafts.
How it compares
Automatic observer memory—not a manual MEMORY.md habit or a full pgvector RAG deployment.
Common Questions / FAQ
Who is total-recall for?
Solo and indie builders who live in agentic IDEs and want durable preferences and project context without maintaining databases or embedding pipelines.
When should I use total-recall?
Install during Build agent-tooling, then rely on it across Validate scoping, Ship review threads, Launch drafting, Grow analytics questions, and Operate firefighting whenever session continuity matters.
Is total-recall safe to install?
Check the Security Audits panel on this Prism page and review what files the observer writes, which models it calls, and whether conversation snippets leave your machine.
SKILL.md
READMESKILL.md - Total Recall
The only memory skill that watches on its own. No database. No vectors. No manual saves. Just an LLM observer that compresses your conversations into prioritised notes, consolidates when they grow, and recovers anything missed. Five layers of redundancy, zero maintenance. ~$0.00/month (using free-tier models). While other memory skills ask you to remember to remember, this one just pays attention. # total-recall { "name": "total-recall", "description": "The only memory skill that watches on its own. No database. No vectors. No manual saves. Just an LLM observer that compresses your conversations into prioritised notes, consolidates when they grow, and recovers anything missed. Five layers of redundancy, zero maintenance. ~$0.00/month (using free-tier models). While other memory skills ask you to remember to remember, this one just pays attention." }