# teamcopilot

> A shared AI Agent for Teams

- **URL**: https://www.freshcrate.ai/projects/teamcopilot
- **Author**: 
- **Category**: AI Agents
- **Latest version**: `0.4.14` (2026-05-25)
- **License**: MIT
- **Source**: https://github.com/rishabhpoddar/teamcopilot
- **Homepage**: https://teamcopilot.ai
- **Language**: TypeScript
- **GitHub**: 9 stars
- **Registry**: npm (`teamcopilot`)
- **Tags**: `npm`

## Description

A shared AI Agent for Teams

## Recent releases

| Version | Date | Urgency | Changes |
| --- | --- | --- | --- |
| `0.4.14` | 2026-05-25 | High | # teamcopilot 0.4.14\n\nThis release adds session-scoped workflow allowlisting so a chat session can opt into always allowing a workflow during that session. The permission prompt and tool-call UI were updated to reflect the new behavior, and the backend now tracks and enforces the session allowlist state.\n\n## Highlights\n- Added a new workflow session allowlist flow for 'always allow in this session'.\n- Updated chat permission handling and tool-call display to surface the new session state.\ |
| `0.4.12` | 2026-05-22 | High | ## Changes - Increased the backend JSON body limit to 10 MB so large file edits can be saved without hitting Express's default 100 KB request cap. - This fixes the 'request entity too large' failure when saving big files from the UI.  ## Release checks - Published to npm as teamcopilot@0.4.12. - The release helper's test suite and build completed successfully before publish. |
| `0.4.0` | 2026-05-16 | High | TeamCopilot v0.4.0  This release is centered on the new cronjobs feature set and the supporting OpenCode integration work behind it.  Highlights: - Added the cronjobs runtime, scheduler, lifecycle management, and resumable chat handoff flow. - Added the cronjob management UI, run pages, and supporting frontend formatting updates. - Added cronjob tooling for managing cronjobs, todos, user handoff prompts, completion, and failure handling. - Added the database migrations and validation logic requi |
| `0.3.6` | 2026-05-02 | High | ## Added - Google Vertex support for managed OpenCode auth in TeamCopilot. - Provider-specific auth guidance for Azure OpenAI and Google Vertex in the setup UI. - Vertex runtime validation and workspace dependency wiring for `@ai-sdk/google-vertex`.  ## Updated - `.env.example` now documents the Vertex environment variables needed for managed setups. - Workspace AGENTS guidance was refreshed to match the current platform description.  ## Notes - Google Vertex-managed setups now require `GOOGLE_C |
| `0.3.5` | 2026-04-30 | High | ## Overview  This release ships the **workflow HTTP API** and supporting persistence so workflows can be triggered outside the web UI with explicit credentials.  ## User-visible changes  - **Workflow API**: Endpoints to run workflows programmatically, aligned with the existing approval and execution model. - **Workflow API keys**: Issue and revoke keys scoped for API access; keys participate in the normal permission and audit boundaries.  ## Upgrade notes  - Existing deployments should run datab |
| `0.3.3` | 2026-04-27 | High | ## What's Changed  This release expands TeamCopilot's shell secret proxy to support top-level git commands, so agents can use declared Profile Secrets for private repository operations without exposing raw token values in chat or command text.  ### Git Secret Injection  - Added secret placeholder rewriting for top-level `git ...` commands. - Supports common private repository workflows such as `git clone`, `git remote set-url`, `git fetch`, `git pull`, `git push`, and `git -c http.extraHeader=.. |
| `0.3.2` | 2026-04-21 | Low | Imported from npm (0.3.2) |
| `0.3.1` | 2026-04-10 | Medium | ## TeamCopilot 0.3.1  This release tightens workflow execution behavior for agent-triggered runs and ships the corresponding package metadata sync needed for reproducible npm releases.  ### Highlights  - `runWorkflow` now requires the agent to pass an explicit `inputs` object instead of silently relying on an omitted argument payload. - The tool-level guidance for `runWorkflow` is clearer, so agents are directed to inspect workflow docs or use `findSimilarWorkflow` to determine the expected argu |
| `0.3.0` | 2026-04-09 | Medium | ## What's new  - Added a token usage dashboard with team-wide estimated usage totals, time-series charts, and model/provider breakdowns - Added incremental session-level usage syncing with estimated cost calculation - Added provider-aware model pricing support plus env-var pricing overrides for the active model - Improved the token usage UI across desktop and mobile layouts - Fixed the npm release script/flow |
| `0.2.2` | 2026-04-06 | Medium | ## What's Changed  ### Fix: bash environment variable injection in secret proxy  The secret-proxy plugin was not reliably injecting resolved secret values as environment variables into bash tool calls. This caused workflows and skills that rely on secrets (e.g. `OPENAI_API_KEY`, `GITHUB_TOKEN`) to fail at runtime because the env vars were empty or missing.  The fix introduces a two-level pending key tracking system:  - **Per-call tracking** — when a `tool.execute.before` event fires for a bash c |

## Citation

- HTML: https://www.freshcrate.ai/projects/teamcopilot
- Markdown: https://www.freshcrate.ai/projects/teamcopilot.md
- Dependencies JSON: https://www.freshcrate.ai/api/projects/teamcopilot/deps

_Generated by freshcrate.ai. Indexes npm releases for AI-agent ecosystem packages._