apiblaze@0.20.39
APIblaze CLI — Chat with your APIs, Manage your API keys, users and groups with the APIblaze serverless proxy
AI Security Review
scanned 2mo ago · by lpm-firewall-aiThe package performs explicit-user-command agent configuration, creating a real capability and persistence risk. Static inspection found no automatic lifecycle trigger, hidden exfiltration sink, or concrete malware behavior.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
The user runs an apichat command with --install-mcp codex or accepts the interactive MCP installation flow.
Impact
Future Codex sessions may gain access to the configured API tools; this is a user-invoked agent capability change.
Mechanism
The CLI writes an MCP server block to the Codex configuration and a marked instruction block to the current directory’s agent instruction file.
AI rationale
The package performs explicit-user-command agent configuration, creating a real capability and persistence risk. Static inspection found no automatic lifecycle trigger, hidden exfiltration sink, or concrete malware behavior.
Evidence
dist/index.js.codex/config.tomlAGENTS.md
Decision evidence
public snapshotAI called this Suspicious at 84.0% confidence as Dangerous Capability with low false-positive risk.
Evidence for warning
- The package offers an explicit option to install an MCP proxy into Codex or Claude.
- The Codex installation routine rewrites the selected MCP block in the user’s Codex configuration.
- The MCP installation flow writes a standing instruction telling agents in the current directory to call the MCP tools.
Evidence against
- The manifest has only a prepublish build/typecheck hook and no install-time lifecycle hook.
- The agent integration is activated by an explicit --install-mcp option rather than package installation or import.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkShellWebSocket
HighEntropyStringsPossibleObfuscationUrlStrings
Source & flagged code
8 flaggeddist/index.jsView file
77const raw = fs.readFileSync(CREDENTIALS_PATH, "utf-8");
L78: return JSON.parse(raw);
L79: } catch {
...
L102: path = __toESM(require("path"));
L103: APIBLAZE_DIR = path.join(os.homedir(), ".apiblaze");
L104: CREDENTIALS_PATH = path.join(APIBLAZE_DIR, "credentials.json");
...
L123: async function createProxyAnonymous(body) {
L124: const res = await fetch(`${PUBLIC_API_BASE}/proxy`, {
L125: method: "POST",
L126: headers: { "Content-Type": "application/json" },
L127: body: JSON.stringify(body)
L128: });
Critical
Credential Exfiltration
Source appears to send environment or credential material to an external endpoint.
dist/index.jsView on unpkg · L7777Trigger-reachable chain: manifest.main -> dist/index.js
L77: const raw = fs.readFileSync(CREDENTIALS_PATH, "utf-8");
L78: return JSON.parse(raw);
L79: } catch {
...
L102: path = __toESM(require("path"));
L103: APIBLAZE_DIR = path.join(os.homedir(), ".apiblaze");
L104: CREDENTIALS_PATH = path.join(APIBLAZE_DIR, "credentials.json");
...
L123: async function createProxyAnonymous(body) {
L124: const res = await fetch(`${PUBLIC_API_BASE}/proxy`, {
L125: method: "POST",
L126: headers: { "Content-Type": "application/json" },
L127: body: JSON.stringify(body)
L128: });
Critical
Trigger Reachable Dangerous Capability
A package entrypoint or install-time lifecycle script reaches a source file with blocking dangerous behavior.
dist/index.jsView on unpkg · L77Findings
3 Critical6 High3 Medium5 Low
CriticalCredential Exfiltrationdist/index.js
CriticalTrigger Reachable Dangerous Capabilitydist/index.js
CriticalPrevious Version Dangerous Deltadist/index.js
HighChild Processdist/index.js
HighShell
HighSame File Env Network Executiondist/index.js
HighCredential Redirect Persistencedist/index.js
HighSandbox Evasion Gated Capabilitydist/index.js
HighTrigger Reachable Credential Exfiltrationdist/index.js
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowNon Install Lifecycle Scripts
Affected versions and remediation
This report applies to apiblaze@0.20.39.
See all versions of apiblaze for other recorded verdicts.
- Review the evidence and your use of apiblaze@0.20.39 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.