Official Claude-Mem demonstration: a coding session beside a browser showing persistent observations.
Captures du projetÀ propos de ce Skill
Persistent memory for coding agents, with 21 curated workflows for recall, code exploration, planning, reviews, reports, and integrations. Install the upstream plugin; hosted memory and model usage may cost extra.
Pourquoi ce choix
Useful when an agent repeatedly loses project decisions between sessions. This collection presents reusable task workflows from the upstream repository. The gallery is the author’s supplied demonstration; it is not an example generated by SkillHey. Separately listed workflows share runtime and helper dependencies.
Skills de cette collection
21 skills- Voir →
Claude-Mem Memory Search
Find previous decisions and solutions with a bounded index → context → selected-details search over Claude-Mem history.
- Voir →
Claude-Mem Smart Explore
Explore a codebase with tree-sitter search, outlines, and selective symbol expansion instead of loading every file.
- Voir →
Claude-Mem Session Handoff
Write a precise HANDOFF.md covering the goal, current state, files, failed attempts, best theory, and next steps.
- Voir →
Claude-Mem Implementation Planning
Build a phased implementation plan grounded in discovered documentation, exact APIs, and verification criteria.
- Voir →
Claude-Mem Execute a Plan
Execute a phased plan with separate implementation, verification, quality review, and branch handoff responsibilities.
- Voir →
Claude-Mem PR Follow-up
Monitor a pull request, address actionable review feedback, and recheck CI until the PR meets its readiness criteria.
- Voir →
Claude-Mem Post-release Verification
Check shipped packages, deployments, installed plugins, and late reports during a defined post-release observation window.
- Voir →
Claude-Mem Design Is
Audit a design against Dieter Rams’ ten principles, score cited evidence, and hand off a concrete design improvement plan.
- Voir →
Claude-Mem Architecture Pathfinder
Map feature flows, identify repeated concerns, and propose a unified architecture with source-linked handoff prompts.
- Voir →
Claude-Mem Issue Backlog Clustering
Cluster GitHub issues by shared root cause, create canonical plan issues, and prepare coherent fix bundles.
- Voir →
Claude-Mem Branch Standup
Compare active worktrees, branches, and PRs in a shared read-only discussion, then produce one consolidation plan.
- Voir →
Claude-Mem Agent Cost Report
Analyze local agent transcripts into an HTML cost report, JSON, CSV line items, and evidence with explicit estimate labels.
- Voir →
Claude-Mem Project Timeline Report
Turn a project’s Claude-Mem timeline into a narrative report of its development, decisions, and debugging history.
- Voir →
Claude-Mem Weekly Project Digests
Split project memory into ISO weeks and generate connected narrative chapters with a carry-forward between weeks.
- Voir →
Claude-Mem Memory Knowledge Agent
Build a focused corpus from project observations, prime an AI session, and query that corpus conversationally.
- Voir →
Claude-Mem Memory Mode Creator
Interview for a useful memory taxonomy, validate and install a custom mode, and verify it in startup context.
- Voir →
Claude-Mem Memory Cloud Sync
Configure or diagnose Claude-Mem SyncHub through the official installer and verify authenticated hub reachability.
- Voir →
Claude-Mem Screenpipe Recall
Search Screenpipe OCR or audio history and optionally preserve selected findings in Claude-Mem.
- Voir →
Claude-Mem Wowerpoint
Send one source document to NotebookLM and start a kawaii narrative slide deck using notebooklm-py.
- Voir →
Claude-Mem Cursor Memory Setup
Set up Claude-Mem in Cursor with a local or remote worker and choose the appropriate observation provider.
- Voir →
Claude-Mem Cowork Memory Pairing
Pair the Claude-Mem Cowork plugin with a personal cmem.ai account and verify its masked configuration status.
À quoi il sert
- Recall past decisions and solutions across coding sessions.
- Explore code, plan changes, coordinate branches, and follow PR or release checks.
- Create project history reports, usage reports, or a focused memory knowledge base.
Capture, compress, recall
Claude-Mem captures agent activity, summarizes it into observations, and supplies relevant context to later sessions. Memory search progressively narrows an index into context and selected details. The local worker also exposes a browser viewer for inspecting observations.
Install the full upstream integration
The main command starts the upstream installer. Choose the runtime and provider appropriate to your setup. In Claude Code, the plugin marketplace commands are another documented route. Restart the host as instructed and start a new session to verify hooks and MCP.
| Commande | À quoi elle sert |
|---|---|
npx claude-mem install | Official installer; defaults to the Claude Code integration. |
/plugin marketplace add thedotmack/claude-mem
/plugin install claude-mem | Run these commands inside Claude Code. |
npx claude-mem install --ide cursor | Cursor integration; other hosts have their own documented --ide choices. |
Official repository and installation guideClaude-Mem documentation
Choose a task from the collection
The 21 members below cover recall, structural code exploration, handoffs, planning and execution, PR/release verification, design and architecture reviews, issue clustering, branch coordination, reports, knowledge corpora, memory configuration, Screenpipe, NotebookLM slides, Cursor setup, and Cowork pairing. Individual entries explain their own inputs, outputs, and additional dependencies.
The upstream inventory has 27 named primary plugin entries plus platform variants. This collection excludes eight primary entries that are repository-specific maintenance, minimal wrappers, explanatory reference, or an insufficient standalone method; it adds two distinct platform setup workflows. The installation bundle still contains upstream files beyond this curated catalog.
Runtime and service choices
Open-source code does not imply all inference or hosting is free. A local setup can use the selected provider or existing agent plan; cmem.ai adds hosted services. Consult the upstream installer and service pricing rather than assuming a permanent free quota. No paid plan is sold through this external listing.
Official repository and installation guideClaude-Mem documentation
Avant de l’utiliser
- Node.js 20+ and a supported agent integration. The local worker uses Bun; some search features use uv/Python dependencies installed by the upstream setup flow.
- Install the actual plugin and its MCP/hooks/helpers. Copying a SKILL.md alone does not provide the memory runtime.
- The code is Apache-2.0. Optional hosted memory, inference providers, agent subscriptions, and NotebookLM services have separate terms and possible charges.
Agents compatibles
Questions fréquentes
Can I install each catalog member by copying only its entrypoint?
Some workflows are instruction-led, but many require the shared worker, MCP tools, scripts, references, or a separate platform bundle. Follow the original installation documentation and each member’s dependency notes.
Is the preview a new result generated for this listing?
No. It is the supplied official GIF, compressed into an animated WebP while retaining the complete duration and loop.
Sources et références
Les sources originales sont liées ci-dessous. Les pistes pratiques sont des suggestions de la rédaction de SkillHey.
