Supermemory Cross-Session Recall
Store, retrieve, and correct project facts and preferences through Supermemory with scoped recall and verified writes.
Before & after
Same prompt, same agent — the only difference is the skill.
That's a common challenge with stateless AI assistants. You might consider keeping a shared document or notes file that you paste into each conversation. Some users maintain a personal wiki with their working style, preferences, and context. You could also structure your prompts to include this information upfront, or use a notes app alongside Claude to track what matters. Would it help if I suggested a note-taking system, or were you looking for something else?
SUPERMEMORY SETUP Automation Status: ACTIVE Memory Mode: Cross-session Stored Items: 127 preferences and contexts YOUR STORED CONTEXT • Writing voice: technical, concise, no jargon • Current projects: AgentVolt integration, API design review • Time zone: Pacific (PDT) • Default tools: Bash, Read, Edit (no web fetch) • Avoid: video/image generation, long explanations RECENT RECALLS ✓ Used your project stack (Next.js, Prisma, Vercel) ✓ Applied your naming convention (kebab-case components) ✓ Referenced your brand guidelines (AgentVolt attribution) LAST SESSION: Yesterday at 3:15pm MEMORY SCORE: 94% accuracy (internal benchmark) To add new preferences, just mention them naturally—Supermemory captures them. To review or clear old memories, use /supermemory-view or /supermemory-clear.
About this skill
“We already decided this” should not mean explaining the whole project again. This skill helps Claude use Supermemory to recall relevant preferences and decisions, correct outdated facts, and keep unrelated projects out of the answer. It also distinguishes a verified memory write from a promise to remember.
What it does
- Checks whether Supermemory is actually connected and which memory operations are available.
- Retrieves relevant facts within the correct user, organization, and project scope.
- Saves durable preferences, decisions, and constraints as individually maintainable entries.
- Handles exceptions and corrections without silently turning a project rule into a global preference.
- Tracks source, effective date, and supersession so old decisions do not masquerade as current facts.
- Verifies writes when readback is available and reports pending or unverified saves accurately.
What you get
Relevant recalled context applied to the task, plus a short receipt when something is saved, corrected, or removed. If Supermemory is unavailable, you get a clearly labeled proposed memory entry rather than a false persistence claim. Installation supplies the workflow; cross-session storage requires an accessible Supermemory integration.
Best for
People using Supermemory with Claude for recurring projects, long-running decisions, and preferences that should carry across conversations without accumulating contradictory notes.
SKILL.md preview
---
name: supermemory-cross-session-recall
description: Maintain cross-session context through an available Supermemory connection using scoped retrieval, atomic fact storage, conflict resolution, and verified write receipts. When no memory connection is actually exposed, say so plainly, do the task from current context, and hand the user a portable carry-forward note. Use when the user says “remember this for next time”, “what did we decide last time?”, “use my saved preferences”, “remind me every session”, “my assistant keeps forgetting the project”, “you keep using an outdated decision”, “I have conflicting saved notes”, “keep these projects separate”, or “forget that saved detail”.
---
… (sign up to view the full skill)More ai & agents skills
View all AI & Agents skills →New Skill Scaffolder
Create a new Claude Code skill correctly: right mechanism, valid name and folder, a triggering description, and a smoke test.
Skill Library Lint and Overlap Report
Lint every skill in a repo at once: frontmatter, name-folder match, weak descriptions, overlapping triggers, broken refs, plus a CI script.
Self-Verifying Agent Loop: Plan, Check, Retry, Close
Turns a goal into a task plan with machine-run checks, executes with retry caps, escalates, and refuses to close until verified.
Skill Test Runner (Trigger and A/B)
Test a skill in fresh sessions: ready-to-run prompts, commands, and rubric for trigger recall, false triggers, and blind A/B.