VESTI Skills
Shared local memory for Codex, Claude Code, Cursor and others: MCP search over past sessions plus a work handoff skill
- 395 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Full agent conversations are copied into a local database and `vesti_summarize` sends selected turns to the configured API.
- The installer edits client configs and starts a background process.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/vesti-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Full agent conversations are copied into a local database and `vesti_summarize` sends selected turns to the configured API.
- The installer edits client configs and starts a background process.
This is third-party code. Review the repository files before installing.
What it does
VESTI Skills collects local session history from Codex, Claude Code, Kimi Code, Cursor, Qoder, WorkBuddy and Trae into one database and exposes it to the agent over MCP. The `vesti-memory` skill tells the client to check the index first and read specific turns only afterwards, so the context is not filled with noise. The second skill, `vesti-handoff`, builds a structured handoff package for another session or model. The receiving agent must rerun the latest verification instead of trusting the package. The `setup` command installs the skills, registers MCP and starts background capture.
Who it is for. Developers who run one project across several agents and are tired of re-explaining context after every session switch.
Good fit when
- A project moved from Codex to Claude Code and you need to recall which option was rejected and why.
- A new session starts with the question of where the last one stopped.
- A scrap of discussion has to be traced to a file to check why it was changed.
- Work is handed to another agent and the handoff package must be verifiable.
Not a fit when
- Session history cannot be collected into a local database, even on your own machine.
- You need cloud memory shared by a whole team.
- Installing from source with pnpm and Node.js 22.12 is too heavy for your case.
Example request
Search past sessions for why we dropped the Redis queue and show the original turns.Limitations
The README says installation currently runs from source: Git, Node.js 22.12 or newer and Corepack are required. The installer writes absolute paths to the build into client configs, so the repository cannot be moved. Conversations from every client accumulate in `~/.vesti`; account for that on private projects. Model summaries through `vesti_summarize` send the selected turns to your API and need checking against the original. The last repository commit is from September 2026.
How to disable. Remove the VESTI skills and MCP entry from client configs and stop the background capture process. Delete the `~/.vesti` directory with the database and logs separately.
FAQ
Which clients are supported?
Codex, Claude Code, Kimi Code, Cursor, Qoder, WorkBuddy and Trae. The README section on supported clients lists the `--host` values.
What should I check after installing?
Run `status`, `sync` and `doctor` from `packages/vesti-memory/dist/cli.js`. `status` should show `Skill=ready, MCP=ready` for your client, and `doctor` ends with `VESTI memory is ready.`.
Can I configure a single client?
Yes. Add the `--host` flag to `setup`, for example `--host codex`. It selects whose config to change and does not limit which sessions are captured.
Is a model required?
No. Without configuration only local search runs. To get `vesti_summarize`, point `~/.vesti/config/llm.json` at an OpenAI Chat Completions compatible API.
Is the VESTI app required?
No. Capture, MCP and skills work independently of the VESTI-APP desktop application, which stays optional.
Discussion
No questions or reviews yet.
Related
SkillsOpen infrastructure for agents that learn from experience: session traces, a knowledge base and skill packs via CLI, MCP and API
31 skills for Claude Code and Codex: cross-vendor review, task handoff on low quota, a morning brief and a nightly retrospective
One local memory for Codex, Claude Code, Cursor, Hermes Agent and OpenClaw: agents hand off tasks, review each other and refine their skills overnight
Local persistent memory for coding agents: hooks record every session, and Claude Code, Codex or Cursor recall past decisions through 54 MCP tools
Shared cloud memory for a team's agents: stores session traces and turns repeated patterns into skills
The xskill CLI gathers sanitized session trajectories from Claude Code, Codex and OpenCode and grows a shared skill library.
README in short
The repository holds an installer CLI, background capture, a read-only MCP server and two skills. The VESTI-APP application is optional.