resume-context

    

resume-context

Never lose context between coding sessions

Ask OpenClaw about any project in plain English · Powered by resume + Redis

---

Never lose context between coding sessions. Ask OpenClaw in plain English from any channel:

"Claw give me a briefing on resume" "Claw what was I working on for my-app?" "Claw show me my notes for the api project"

You get back a real, LLM-generated summary of what you were working on, what you finished, and what to do next — sourced directly from your local session data, not a memory system.

---

Demo

Session briefing — "Claw give me a briefing on resume"

Project notes — "Claw show me my notes for resume"

resume running in VS Code — the TUI dashboard alongside your code

---

How it works

You
 │  "Claw give me a briefing on resume"
 ▼
OpenClaw
 │  skill triggered by message pattern
 ▼
resume-context
 │  cache key: resume:show:/path/to/project
 ▼
Redis
 ├── HIT  → return instantly (<100ms)
 └── MISS → run `resume show`
              │  calls Anthropic LLM (~12s)
              └─ write to Redis, TTL 5 min
 ▼
OpenClaw
 │  "Resume Project Briefing:
 │   You're updating docs to showcase the TUI dashboard.
 │   Progress: Updated README, replaced VSCode screenshot.
 │   Next: Verify ClawHub integration end-to-end."
 ▼
You

First request: ~12 seconds (LLM call via resume). Every repeat in the next 5 minutes: under 100ms (Redis cache hit).

---

Install

openclaw skills install resume-context

Add to ~/.openclaw/openclaw.json:

{
  "skills": {
    "entries": {
      "resume-context": {
        "enabled": true,
        "env": {
          "REDIS_URL": "redis://default:password@host:port",
          "RESUME_CACHE_TTL": "300"
        }
      }
    }
  }
}

Restart the gateway:

openclaw gateway restart

---

Requirements

---

Usage

Start a resume session in your project before asking:

cd ~/your-project
resume          # starts watching your session

Then in OpenClaw:

---

Files

resume-context/ ├── SKILL.md — OpenClaw skill definition: trigger phrases + agent instructions ├── resume-mcp.js — Node.js bridge: Redis cache layer + resume CLI runner └── package.json — Single dependency: redis ^4.7.0

SKILL.md teaches OpenClaw when to trigger this skill and exactly how to run it. The description frontmatter is injected into the model's system prompt — that's what makes "Claw give me a briefing on resume" route here instead of a generic memory search.

resume-mcp.js is the Redis caching bridge. It checks the cache first, falls back to the live resume CLI on a miss, writes the result back with a configurable TTL, and returns structured JSON. If Redis is unreachable it degrades gracefully to live execution.

---

Environment variables

---

License

MIT

---

   

A skill for OpenClaw · Built at the OpenClaw Agent Toolkit Hack Day · March 25, 2026 · San FranciscoSponsors: Redis · OpenAI Codex · HackerSquad