Screenpipe Chats

Search what was said in existing screenpipe, Codex, Claude, Cursor, and Gemini CLI chats, then continue or steer one exact chat when the user explicitly asks. Use for requests about finding another agent conversation, recalling what an agent said, or sending work to one.

22,429 stars 2,305 forksAI, LLMs & Agents

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add screenpipe/screenpipe --skill screenpipe-chats
Manual install
git clone --depth 1 https://github.com/screenpipe/screenpipe.git /tmp/screenpipe
mkdir -p ~/.claude/skills
cp -r /tmp/screenpipe/crates/screenpipe-core/assets/skills/screenpipe-chats ~/.claude/skills/screenpipe-chats

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# existing agent chats

screenpipe exposes `search_chats` and `send_to_chat` as agent tools. They use a
private core capability passed into the current chat process; do not scrape
transcript folders, guess runtime commands, or call a localhost app route.

## search

1. Call `search_chats` with a short title, message, workspace, or exact id. The
   query also matches the transcript, so a phrase said mid-conversation finds
   the chat even when the title never mentions it. Omit the query to list recent
   chats. Filter `sources` only when the user named a runtime.
2. Show enough title, source, preview, and workspace information to distinguish
   ambiguous results.
3. Treat the returned `source` and `id` as the address. Never derive an id from
   a title or reuse a stale result after the target disappears.
4. Read the `warnings` array. Transcript reading is bounded to the most recent
   chats per runtime, so report that older chats were not searched instead of
   claiming a phrase was never said.

Searching is read-only and does not require confirmation.

## send

1. Require an exact result from `search_chats` in the current turn.
2. Require explicit user authorization for that target and message. A draft,
   suggestion, or request to inspect chats is not permission to send.
3. Call `send_to_chat` with the exact `source`, `id`, message, and
   `confirmed: true`.
4. Use `mode: queue` by default. Use `mode: steer` only when the user asked to
Read the full skill on GitHub

Details

Skill name
screenpipe-chats
Path
crates/screenpipe-core/assets/skills/screenpipe-chats/SKILL.md
Repo updated
17 November 2026

FAQ

What does the Screenpipe Chats skill do?

Search what was said in existing screenpipe, Codex, Claude, Cursor, and Gemini CLI chats, then continue or steer one exact chat when the user explicitly asks. Use for requests about finding another agent conversation, recalling what an agent said, or sending work to one.

How do I install the Screenpipe Chats skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add screenpipe/screenpipe --skill screenpipe-chats. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Screenpipe Chats popular?

It lives in screenpipe/screenpipe, which has 22,429 GitHub stars and 2,305 forks, and is included in 1 curated skill list.

Related skills

Develop Screenpipe Windows

22k

screenpipe/screenpipe

Develop and test Screenpipe Windows-native changes on a disposable Azure VM created from the prepared Screenpipe Windows dev image. Use for Windows compiler, process, service, local API, desktop, capture, installer, or permission behavior. Do not use for React-only work proved by the browser mock.

Software DevelopmentIn 1 list2.3k

Release

22k

screenpipe/screenpipe

Release the screenpipe monorepo. Bumps versions, triggers GitHub Actions for app, CLI, MCP, and JS packages.

Git & Code ReviewIn 1 list2.3k

Render HTML Report

22k

screenpipe/screenpipe

Produce a human-facing visual report — a chart, dashboard, scorecard, or styled summary — that renders as a live page in the screenpipe viewer instead of plain text. Use when the task asks for a visual/graphical output rather than a plain note; do NOT use for plain text or raw data (prefer a markdown note for those).

Productivity & NotesIn 1 list2.3k

Plannotator Annotate

13k

backnotprop/plannotator

Open Plannotator's annotation UI for a markdown file, HTML file, URL, or folder and then respond to the returned annotations.

AI, LLMs & AgentsIn 2 lists1.3k

Gws Modelarmor Sanitize Response

32k

googleworkspace/cli

Google Model Armor: Sanitize a model response through a Model Armor template.

AI, LLMs & AgentsIn 2 lists2k

Plannotator Compound

13k

backnotprop/plannotator

Analyze a user's Plannotator plan archive to extract denial patterns, feedback taxonomy, evolution over time, and actionable prompt improvements — then produce a polished HTML dashboard report. Falls back to Claude Code ExitPlanMode denial reasons when Plannotator data is unavailable.

AI, LLMs & AgentsIn 2 lists1.3k

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.