@highflame/overwatch
IDE-agnostic security daemon for AI coding assistants. Evaluates prompts, tool calls, file operations, shell commands, and MCP usage in real time for Cursor, Claude Code, GitHub Copilot, Gemini CLI, OpenAI Codex, and Kiro.
This report shows the evidence for @highflame/overwatch@2.1.3. 0 of 8 known versions are blocked or flagged. See all versions.
AI Security Review
scanned 3d ago · by lpm-firewall-aiThe daemon automatically sends authenticated users’ email addresses and IDE identity to Highflame during hook processing, with no explicit telemetry opt-in found in the inspected flow. This is identifying runtime telemetry and warrants a warning; the package’s prompt and tool-event evaluation traffic otherwise matches its stated purpose.
Decision evidence
public snapshot- The daemon records the authenticated user’s email as both user_id and user_email, along with the IDE type.
- The hook pipeline registers a callback that checks installation status; when authenticated with an email, it automatically submits an installation record.
- Installation records are POSTed to the configured Highflame API, whose default host is control.highflame.ai.
- The daemon also sends hook event payloads to its Cerberus service, consistent with the package’s stated prompt and tool-call evaluation purpose.
- Installation reporting is limited to authenticated users with an available email.
- Evaluating AI assistant prompts and tool calls is the package’s stated purpose, and those event payloads are sent to its own service.
Source & flagged code
8 flaggedPackage source references dynamic require/import behavior.
dist/index.jsView on unpkg · L10Findings
Affected versions and remediation
This report applies to @highflame/overwatch@2.1.3.
The versions list below shows the verdict for each recorded version.
- Review the evidence and your use of @highflame/overwatch@2.1.3 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.