Run a FlareDeck agent on your own machine: it picks up cards assigned to it, works each one in a checkout of its own, and reports back on the card.
Flagged — allowed with a warning
Allowed by default policy, but 19 findings warrant review before installing.
AI security review
scanned · by lpm-firewall-ai
The package provides an undisclosed, default-on remote shell controlled by the workspace manager, with no per-command local approval. The manager and service are part of the package’s intended workflow and a local opt-out exists, so this warrants a warning rather than a malicious verdict.
Static reason
One or more suspicious static signals were detected.
Trigger
The agent is running and receives an `exec` event from its workspace manager.
Impact
A workspace manager can run commands with the agent user’s local permissions and receive their output, without a separate per-command local approval.
Mechanism
The handler claims the request using the agent’s authentication, runs the command with `bash -lc` as the local user, and posts the result to the configured FlareDeck origin. A local `remote off` file blocks execution.
Evidence
dist/remote-exec.jsdist/serve.jsREADME.md
Network endpoints1
Configured FlareDeck MCP origin
Decision evidence
public snapshot
Evidence for warning
A manager’s exec event can reach a handler that claims the command and runs it locally with bash.
The handler posts command output to the configured FlareDeck service.
Remote commands are enabled unless the user creates a local opt-out file; the README does not disclose this feature.
The package has no install, preinstall, or postinstall script; its prepublishOnly script builds the package.
Evidence against
The remote command request comes from the workspace manager through the configured FlareDeck service.
A local `flaredeck-agent remote off` switch refuses requests, and FlareDeck cannot remove that switch.
Running assigned coding tasks is the package’s stated purpose, and its normal agent work uses command execution.
Behavioral surface
SourceChild process · Crypto · Environment variables · Eval · Filesystem · Network · Shell · Web socket
Supply chainHigh entropy strings · URL strings
ManifestNo license
Scanned 130 files · 1.58 MB of source
14 external domains referenced
127.0.0.1
api.github.com
app.flaredeck.app
flaredeck-agent-861259873190.us-central1.run.app
generativelanguage.googleapis.com
github.com
jam.dev
mcp.cnjure.design
mcp.flaredeck.app
mcp.jam.dev
mcp.sceneflo.ai
openrouter.ai
pypi.org
registry.npmjs.org
Findings
6 high · 5 medium · 8 low
Where they are, most severe file first. Code opens a file's flagged lines, with their evidence, below.