ft-scout@9.5.2
FrontTerrain — The Onboarding Co-Pilot for New Codebases
AI Security Review
scanned 18h ago · by lpm-firewall-aiThe package’s agent feature legitimately sends user prompts to an AI service, and the telemetry upload is gated on a companion session ID. However, it also streams prompt text and project names to Firestore, with no clear per-session disable control, so this is a runtime telemetry concern rather than clear malicious exfiltration.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
A user runs the agent while a sessionId remains in ~/.ft/scout_telemetry.json from a paired session.
Impact
Task text and project identifiers may be disclosed to the package operator’s companion telemetry service beyond the AI task request itself.
Mechanism
Agent execution writes its goal, project basename, and activity details into the local telemetry snapshot; the telemetry helper merges the snapshot and PATCHes it to Firestore when it finds a sessionId.
AI rationale
The package’s agent feature legitimately sends user prompts to an AI service, and the telemetry upload is gated on a companion session ID. However, it also streams prompt text and project names to Firestore, with no clear per-session disable control, so this is a runtime telemetry concern rather than clear malicious exfiltration.
Evidence
bin/src/utils/qrBridge.jsbin/src/engine/agentEngine.jsbin/src/commands/agent.js~/.ft/scout_telemetry.json
Network endpoints1
firestore.googleapis.com
Decision evidence
public snapshotAI called this Suspicious at 84.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- The telemetry helper sends session snapshots to Firestore at firestore.googleapis.com.
- Cloud upload occurs when a sessionId is present in the saved telemetry snapshot.
- Agent execution records the full task goal, project basename, and a goal excerpt in telemetry.
- The agent engine sends telemetry when an execution starts, so this can recur during normal agent use.
Evidence against
- The cloud upload is session-gated, consistent with a previously paired Scout Companion session.
- The package describes itself as an AI coding assistant, and sending a user’s task to an AI service is part of its purpose.
- The QR pairing flow tells the user that telemetry streams to Scout Mission Control.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkShell
HighEntropyStringsUrlStrings
Source & flagged code
15 flaggedbin/src/utils/firebaseAuth.d.tsView file
2patternName = google_api_key
severity = high
line = 2
High
High Secret
Package contains a high-severity secret pattern.
bin/src/utils/firebaseAuth.d.tsView on unpkg · L22patternName = google_api_key
severity = high
line = 2
High
Secret Pattern
Google API key in bin/src/utils/firebaseAuth.d.ts
bin/src/utils/firebaseAuth.d.tsView on unpkg · L2bin/src/server/server.jsView file
3import path from 'path';
L4: import { exec } from 'child_process';
L5: import { fileURLToPath } from 'url';
High
Child Process
Package source references child process execution.
bin/src/server/server.jsView on unpkg · L3Findings
4 Critical11 High3 Medium5 Low
CriticalHardcoded Runtime Data Exfiltrationbin/src/utils/qrBridge.js
CriticalCredential Exfiltrationbin/src/utils/email.js
CriticalTrigger Reachable Dangerous Capabilitybin/src/utils/qrBridge.js
CriticalPrevious Version Dangerous Deltabin/src/engine/computer/computerController.js
HighHigh Secretbin/src/utils/firebaseAuth.d.ts
HighChild Processbin/src/server/server.js
HighShellbin/src/commands/watch.js
HighSame File Env Network Executionbin/src/utils/holoseneAuth.js
HighSandbox Evasion Gated Capabilitybin/src/commands/setup.js
HighRemote Agent Bridgebin/src/engine/agentEngine.js
High
Affected versions and remediation
This report applies to ft-scout@9.5.2.
See all versions of ft-scout for other recorded verdicts.
- Review the evidence and your use of ft-scout@9.5.2 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.