# Share

Turn a session into a team learning: summarize what was solved, discovered or worked...

- Repository: https://github.com/Tencent/teamai-cli (5729 stars)
- Category: Engineering
- License: see repository
- Last updated: 2026-11-01
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add Tencent/teamai-cli --skill share -g -a claude-code -y`
Codex: `npx skills add Tencent/teamai-cli --skill share -g -a codex -y`
Cursor: `npx skills add Tencent/teamai-cli --skill share -g -a cursor -y`
Gemini CLI: `npx skills add Tencent/teamai-cli --skill share -g -a gemini-cli -y`

## Description

Turn a session into a team learning: summarize what was solved, discovered or worked around, and publish it to the team knowledge base with `teamai contribute`. Loaded on demand by the teamai discovery stub, by the learning step of the teamai-share-learnings workflow, and by the friction reminder where teamai-share-learnings is not offered.

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

# Contribute, share what a session taught you with the team

Summarize what this AI coding session taught you and push it to the team knowledge base.

## Who sent you here

- **The `learning` step of a `teamai-share-learnings` run** (its brief names this skill): work
  steps 1, 3 of How It Works with the person, in this session rather than a sub-agent. The session
  may have taught more than one thing: write each learning the person wants to share to its own
  document outside the repository, and return them as the list the brief asks for, each with its
  path, title and learnings namespace; return an empty list when nothing is worth sharing. Do not
  run `teamai contribute`: the run's `contribute` step runs it for each one after its approval.
- **The friction reminder** (its `[teamai]` text names `teamai skill get share`): teamai already
  found that `teamai-share-learnings` cannot run here. Follow How It Works to the end.
- **The person asked to share what this session taught, or to close it**: start the
  `teamai-share-learnings` workflow (to only publish or score it, `teamai-share-session`). Run
  `teamai workflow start --describe teamai-share-learnings --driver skill`,
  give the person each proposed input with its `why`, and start it with those values once they
  confirm or change them (`teamai skill get workflow` covers the run). Its `learning` step brings
  you back here. When describe refuses it (`invoke_denied`, `unattended_denied`,
  `nested_workflow`), tell the person why and follow How It Works to the end.
- **Anything else in When to Use**: follow How It Works to the end.

**Write the document in Simplified Chinese**, as earlier releases required. Commands, flags,
URLs, paths and code identifiers stay as they are.

## When to Use

The rest: https://github.com/Tencent/teamai-cli/blob/HEAD/skill-data/share/SKILL.md
