@sentropic/h2a-runtime@0.93.4
h2a heavy runtime (sessions / k8s / tunnel, node-pty + aws-sdk). Lazily imported by @sentropic/h2a; installed separately.
AI Security Review
scanned 3mo ago · by lpm-firewall-aiThis package has dangerous, user-invoked agent extension setup capabilities, including broad local agent configuration writes. The behavior is command-gated and package-aligned, so it does not meet the block threshold for malicious install-time control-surface mutation.
Static reason
One or more suspicious static signals were detected.
Trigger
User runs the package's `plugin add` or `plugin sync` command.
Impact
Configured MCP plugins can execute in Claude, Codex, or Gemini/agy environments; arbitrary installer commands require explicit user configuration.
Mechanism
User-directed plugin installation and MCP configuration mutation.
AI rationale
This package has dangerous, user-invoked agent extension setup capabilities, including broad local agent configuration writes. The behavior is command-gated and package-aligned, so it does not meet the block threshold for malicious install-time control-surface mutation.
Evidence
package.jsondist/plugin.jsdist/index.jsdist/auth-tools.jsdist/llm-mesh.js~/.claude.json~/.codex/config.toml~/.gemini/config/mcp_config.json~/.config/sentropic/remote.json~/.remote-manifest.json~/.remote-manifest.sha256
Decision evidence
public snapshotAI called this Suspicious at 92.0% confidence as Dangerous Capability with low false-positive risk.
Evidence for warning
- Explicit `plugin add` installs packages and registers MCP servers.
- Plugin registration writes Claude, Codex, and Gemini agent configuration.
- A user-provided installer script can be persisted and later run in remote Pods.
Evidence against
- package.json has no preinstall, install, or postinstall hook.
- Agent-config and installer actions are reachable through explicit plugin commands.
- No unconsented exfiltration, destructive action, or fixed external credential gateway was confirmed.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkShell
HighEntropyStringsUrlStrings
Source & flagged code
5 flaggeddist/index.jsView file
•Manifest-reachable source captures an API credential, sends it to a fixed unofficial gateway, and persists that redirection in agent or shell configuration.
dist/index.js:
import { existsSync, mkdirSync, readdirSync, readFileSync, realpathSync, statSync, writeFileSync, } from "node:fs";
import { createInterface } from "node:readline";
baseUrl: opts.remote,
async function refreshProfileSession(baseUrl, sessionId, opts) {
const remoteProfile = (await getRemoteSession(baseUrl, sessionId)).session
const response = await refreshRemoteSession(baseUrl, sessionId, credentials, fetch, opts.name);
async function pushAllProfiles(baseUrl, sessionId, opts) {
process.stderr.write(`[h2a] sending creds for ${sent.join(", ")} to ${baseUrl}/sessions/${sessionId}: ${Object.keys(merged).join(", ")}\n`);
High
Credential Redirect Persistence
Manifest-reachable source sends a prompted API credential to a fixed unofficial gateway and persists the redirection.
dist/index.jsView on unpkgdist/pty.jsView file
76package = @sentropic/h2a-runtime; repositoryIdentity = h2a; dependency = node-pty
L76: try {
L77: require("node-pty");
L78: }
High
Copied Package Dependency Bridge
Package metadata claims a different repository identity while copied source loads a runtime dependency bridge.
Findings
3 High5 Medium3 Low
HighCredential Redirect Persistencedist/index.js
HighCopied Package Dependency Bridgedist/pty.js
HighExternal Ai Agent Control Surface Mutationdist/plugin.js
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
MediumAi Review Evidencedist/plugin.js
MediumAi Review Evidencedist/plugin.js
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
Affected versions and remediation
This report applies to @sentropic/h2a-runtime@0.93.4.
See all versions of @sentropic/h2a-runtime for other recorded verdicts.
- Review the evidence and your use of @sentropic/h2a-runtime@0.93.4 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.