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.
Allowed — no malicious behavior detected
20 low-signal patterns reviewed and cleared.
AI security review
scanned · by lpm-firewall-ai
The flagged command-output flow is an intended feature of this FlareDeck agent: it claims manager requests through the authenticated service, runs them, and returns results there. Remote commands are enabled by default, but the package exposes a local remote off switch; the inspected source shows no unrelated collection or concealed execution.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
A FlareDeck `exec` event while the agent is running, after its request is claimed.
Impact
A manager can run commands with the agent user's privileges, which is a broad but product-aligned capability.
Mechanism
Runs the claimed command through `bash -lc` in the agent repository and reports its output to the configured FlareDeck API; a local switch can disable this feature.
Decision evidence
public snapshot
Evidence for AI clean decision
The manifest describes an agent that picks up assigned cards and reports back.
Remote execution is triggered by a FlareDeck `exec` event and uses the agent's authenticated FlareDeck API.
A claimed manager command runs in the agent's repository and its output is returned to FlareDeck.
The CLI exposes a local `remote off` switch that prevents remote commands.
Evidence against
Remote commands are ON by default and can run arbitrary shell commands as the agent's user without a local per-command confirmation.
Command output is sent to the configured FlareDeck service.
Behavioral surface
SourceChild process · Crypto · Environment variables · Eval · Filesystem · Network · Shell · Web socket
Supply chainHigh entropy strings · URL strings
ManifestNo license
Scanned 132 files · 1.63 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
1 critical · 6 high · 5 medium · 8 low
Automated scan signals; the review checked them and cleared this version. Where they are, most severe file first. Code opens a file's flagged lines, with their evidence, below.
382 { type: 'text', text: `This is the screen at ${viewport.width}×${viewport.height}. Judge only what is plain to see; never give a size in pixels.${takeStuckNote()}` },
31// npx puts npm[redacted] in the environment, and the environment beats the config FILE
32// we just wrote — which is why `npm config set prefix` appeared to work and the install
⋯
58functionshellProfile() {
59const home = homedir();
60const shell = process.env['SHELL'] ?? '';
61const candidates = shell.includes('zsh')
62 ? ['.zshrc', '.zprofile', '.profile']
63 : ['.bashrc', '.bash_profile', '.profile'];
'ssh -o BatchMode=yes'
'never'
35});
⋯
44}
45/** `https://github.com/Owner/Repo(.git)` → `owner__repo`: the clone's folder name. */