# onecli/onecli

Open-source sandboxed agent harness for teams. Giving every employee a secured personal agent.

Repository: https://github.com/onecli/onecli
Canonical: https://ross.abutalabs.com/products/onecli
Homepage: https://onecli.sh
Language: TypeScript
License: Apache-2.0
License Family: permissive
Topics: ai-agents, rust, secret-management, cli, nodejs, postgres, security, security-tools, vault, mcp
Last push: 2026-08-26T06:34:38+00:00

## Health v2 (maintenance only)
Score: 77/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 87, longevity 12
- inputs: {"age_days": 178, "days_push": 7, "days_rel": 7, "gap_med": 0, "n_releases_24m": 96}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3415, forks 206 (observed 2026-08-28T04:08:03.279534+00:00)

## What it is
OneCLI is an open-source platform that gives every employee a personal, sandboxed AI agent, with all tool calls routed through a gateway that injects scoped credentials and enforces team policy. Agents never see raw secrets, and human-in-the-loop approvals, IdP provisioning, and Slack access are built in.

## Use cases
- give every employee a secure personal AI agent
- run AI agents without exposing API keys to the model
- enforce a single security policy across all team agents
- provision agents from a company identity provider
- require human approval for sensitive agent actions
- self-host a multi-user agent platform
- route agent tool calls through a credential-injecting gateway

## When to choose
- you want to deploy agents across a team or company with per-person isolation
- your compliance posture requires that agents and LLMs never hold real credentials
- you need deterministic, network-level policy enforcement rather than prompt-based guardrails
- you want IdP integration, audit logs, and Slack access out of the box

## When to avoid
- you are a solo user who just wants a personal agent with full machine access
- you need a lightweight agent framework library rather than a hosted/self-hosted platform
- you cannot run a gateway in front of your agents' network traffic

## Facets
- artifact type: service
- maturity: active
- function: agent-framework, secrets-management, security, mcp, api-gateway
- domain: security, developer-tools, self-hosted
- platform: self-hosted, rust, cli
- tags: credential-vault, sandboxed-agents, human-in-the-loop, team-agents, policy-enforcement, identity-provider, slack-integration, gateway-proxy, sandboxing, ai-agents, docker, nodejs, web-server

## Member repositories
- onecli/onecli (main) score 77

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:03.279534+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:38:28.799269+00:00, confidence not recorded.
  - readme: https://github.com/onecli/onecli (fetched 2026-08-28T04:08:03.279534+00:00, sha ab2cb371de0f)
  - homepage: https://onecli.sh (fetched 2026-08-29T09:32:26.431658+00:00, sha f429d061204e)
  - site_page: https://onecli.sh/docs (fetched 2026-08-29T09:32:26.445087+00:00, sha 113503827ec1)
  - site_page: https://onecli.sh/pricing (fetched 2026-08-29T09:32:26.441024+00:00, sha 0924a9d01121)
  - site_page: https://onecli.sh/faq (fetched 2026-08-29T09:32:26.443277+00:00, sha 0d56602f9009)
- Data as of 2026-08-30T08:39:29.467469+00:00.