Local-first observability for AI coding agents. Every Claude Code and Codex session's tokens, cache, models and cost, on this machine and every machine you connect through a self-hosted team hub. Presenting mode, policy hooks, Apple-signed and notarized macOS builds, attested releases. MIT.
# Add to your Claude Code skills
git clone https://github.com/LockedinLabs-AI/agent-consoleGuides for using ai agents skills like agent-console.
Last scanned: 9/28/2026
{
"issues": [],
"status": "PASSED",
"scannedAt": "2026-09-28T10:37:06.188Z",
"npmAuditRan": true,
"pipAuditRan": true,
"promptInjectionRan": true
}See how agent-console compares with popular alternatives.
agent-console is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by LockedinLabs-AI. Local-first observability for AI coding agents. Every Claude Code and Codex session's tokens, cache, models and cost, on this machine and every machine you connect through a self-hosted team hub. Presenting mode, policy hooks, Apple-signed and notarized macOS builds, attested releases. MIT. It has 803 GitHub stars.
Yes. agent-console passed SkillsLLM's automated security scan — a dependency vulnerability audit plus prompt-injection heuristics — with no high-severity issues. You can read the full report in the Security Report section on this page.
Clone the repository with "git clone https://github.com/LockedinLabs-AI/agent-console" and add it to your Claude Code skills directory (see the Installation section above).
agent-console is primarily written in JavaScript. It is open-source under LockedinLabs-AI on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh agent-console against similar tools.
No comments yet. Be the first to share your thoughts!
⚠️ Third-Party Software Notice
This skill is third-party open-source software developed and hosted independently on GitHub. SkillsLLM is an informational directory and does not control or maintain the underlying repository.
Any security checks, ratings, or warnings displayed by SkillsLLM are automated and limited in scope. They do not constitute a security certification or guarantee that the software is safe, error-free, or free from malicious code, vulnerabilities, compromised dependencies, or prompt-injection risks.
Review the source code, permissions, dependencies, and configuration before installing or running any third-party skill. Use is at your own risk. To the maximum extent permitted by applicable law, SkillsLLM is not liable for losses arising from third-party software.
Captured from --demo. Every figure in it is generated and stamped DEMO.
The same screen in the light theme.
Website · Get started · Documentation · Architecture · Security · Contribute · Changelog
| Local-first | Reads the agents' own transcripts on your machines. No telemetry, no update check, no crash reporting — see data flows. |
| Every machine, one view | Join laptops, build boxes and servers to a self-hosted team hub; each machine keeps its own lanes and nothing is counted twice. |
| Gateway and telemetry aware | Optional ingest of Claude Code OpenTelemetry and Kong or LiteLLM token metrics, a Prometheus /metrics endpoint and a Grafana dashboard. |
| Honest numbers | Estimates say they are estimates; unknown readings are shown as unknown, never as zero. |
| Safe to present | Presenting mode (P) replaces every project, machine and person with a stand-in name. |
| Verifiable supply chain | Apple-signed and notarized macOS builds, SHA256SUMS, signed build attestations and, from 0.4.1, a CycloneDX SBOM on every release (verify one); threat model, NIST SSDF mapping and an enterprise security FAQ. |
Installing with Claude Code or Codex? Give it this repository's URL and ask it to follow INSTALL.md. The same guide works if you prefer to install it yourself: choose the current source, an npm archive, or a standalone download, then verify the version and open the console.
The Accounts view adds native account capacity alongside the token console: remaining allowance, reset times, fresh-data checks, and profile selection for new Codex or Claude Code sessions. Connect accounts and understand the readings. Native sign-in stays with the provider; Agent Console does not pool subscription credentials or proxy model requests.
This page describes v0.4.4. Download a versioned package from GitHub Releases, or run the current source:
git clone https://github.com/LockedinLabs-AI/agent-console.git
cd agent-console
node bin/agent-console.mjs --open
You need Git and Node.js 22 or newer, or use Download ZIP
instead of Git. There is no account to create, no package install and no build
step. The command runs this checkout, reads the Claude Code and Codex history already on
this computer, and opens the console in your browser, signed in, normally at
http://127.0.0.1:6787. To look around first without reading anything of
yours, add --demo before --open.
In the first thirty seconds you see: the last 24 hours of tokens (or the last hour, 7 days or 30 days), split into cache read, cache write, output and uncached input; their list-price estimate; burn right now, in tokens per minute and dollars per hour; one lane per session with its model, its subagents and its last hour of activity; and every machine and person reporting, each with a share of the total.
Nothing leaves a machine but counts. Only model ids, minute timestamps,
token counts and salted hashes, and never a prompt, a reply, a file path or a
file's contents. A test pushes transcripts full of planted canaries through
the real reporter and the real hub and checks every byte that crosses the wire.
The console itself sends nothing anywhere: no telemetry, no update check, no
account. (Three opt-ins, each off unless you pass it: --share-project-names
sends each project folder's name, never its path; --share-alerts and
--share-tool-activity send alerts and tool calls as kinds and counts.)
The installation guide covers all three routes and what to do after closing the console. Downloads live in GitHub Releases. There is no separate desktop installer or hosted dashboard to sign up for; the executable starts the console in your browser.
Choose a version as well as an installation method:
| Channel | Current availability |
|---|---|
| Source on main | v0.4.4 source. Use Node.js 22+ and the commands above, or Download ZIP below. |
| GitHub release downloads | Versioned packages for Node.js and standalone files for macOS, Linux and Windows. Check the listed version and Assets before installing. |
| npm registry | @lockedinlabs/agent-console is not published. Use the source checkout or a published GitHub release archive. |
| Homebrew | The documented tap has not been verified as available. Use source or the listed GitHub release. |
| Published container image | The Linux team hub image v0.4.3 is published for amd64 and arm64. Use the versioned image and verification instructions. |
Prefer npm install? npm can install a published GitHub archive even while
the registry name is unavailable. To install v0.4.4 once its archive is
listed on the release page:
npm install --global --ignore-scripts https://github.com/LockedinLabs-AI/agent-console/releases/download/v0.4.4/lockedinlabs-agent-console-0.4.4.tgz
agent-console --open
In Windows PowerShell use npm.cmd and agent-console.cmd.
If npm reports a permission error, use the source route; administrator access
is unnecessary.
These commands require lockedinlabs-agent-console-0.4.4.tgz to be listed
on the release page.
Until that file is published, use the source instructions above.
npx --yes https://github.com/LockedinLabs-AI/agent-console/releases/download/v0.4.4/lockedinlabs-agent-console-0.4.4.tgz --open
In Windows PowerShell, use npx.cmd because the default policy can refuse
npx's script form:
npx.cmd --yes https://github.com/LockedinLabs-AI/agent-console/releases/download/v0.4.4/lockedinlabs-agent-console-0.4.4.tgz --open
Published files can be checked against the release's SHA256SUMS and build
attestation (Verify a release). Consult that
version's release notes for its signing status. The Windows executable is
not code-signed; Node.js can run the source without installing that executable.
Behind a proxy or TLS inspection. npx, npm and curl use
HTTPS_PROXY; Node.js's own downloads, such as the check at the start of a
join command, use it only with NODE_USE_ENV_PROXY=1 as well (Node.js 22.21
or newer). Where the network inspects TLS, point Node.js at your company's
root certificate with NODE_EXTRA_CA_CERTS=<file>.pem. A join command that
stops with fetch failed on a company network needs these
(details).
To remove it, see Uninstall.
You need a Mac, a Linux machine or a Windows PC, and about two minutes.
node --version — it shou