registry / @exulu/backend / 3.6.0
@exulu/backend@3.6.0
**Exulu IMP** (Intelligence Management Platform) is an open-source TypeScript framework for building production-ready AI agent applications. It provides the backend infrastructure for managing agents, semantic search contexts, background job processing, a
AI Security Review
scanned 3mo ago · by lpm-firewall-aiSource confirms a broad, user-invoked AI-agent skill deployment mechanism, but not unconsented lifecycle execution, exfiltration, or stealth persistence. The scanner's malicious label overstates the observed behavior.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
A user downloads the bootstrap and invokes its helper's install or update command.
Impact
Can modify configured project or home AI-agent skill directories and fetch skill bundles from the configured backend.
Mechanism
configured multi-client skill copy/symlink fan-out
AI rationale
Source confirms a broad, user-invoked AI-agent skill deployment mechanism, but not unconsented lifecycle execution, exfiltration, or stealth persistence. The scanner's malicious label overstates the observed behavior.
Evidence
package.jsonscripts/postinstall.cjsdist/index.js.agents/skills.claude/skills.windsurf/skills.continue/skills
Decision evidence
public snapshotAI called this Suspicious at 88.0% confidence as Dangerous Capability with low false-positive risk.
Evidence for warning
- Runtime code defines a bootstrap manifest for many AI-agent skill directories.
- The generated bootstrap explicitly supports installing or updating skills into agent clients.
- A runtime route generates and serves the bootstrap ZIP.
Evidence against
- npm preinstall only enforces a Node version.
- npm postinstall runs the package-local Python environment setup, not the agent bootstrap.
- No source evidence of automatic agent-directory mutation, credential exfiltration, or foreign endpoint hard-coding during install.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkShell
HighEntropyStringsUrlStrings
NoLicenseWildcardDependency
Source & flagged code
12 flaggedpackage.jsonView file
•scripts.preinstall = node -e "if (process.version !== 'v22.18.0') { console.error('❌ Wrong Node.js version. Expected v22.18.0, got ' + process.version); process.exit(1); }"
Critical
Red Install Lifecycle Script
Install-time lifecycle script matches a deterministic static-gate block pattern.
package.jsonView on unpkg•scripts.preinstall = node -e "if (process.version !== 'v22.18.0') { console.error('❌ Wrong Node.js version. Expected v22.18.0, got ' + process.version); process.exit(1); }"
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg•scripts.postinstall = node scripts/postinstall.cjs
Medium
Ambiguous Install Lifecycle Script
Install-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgbin/backend.cjsView file
9L10: const { spawn } = require('child_process');
L11: const { existsSync } = require('fs');
High
dist/index.jsView file
23356import { fileURLToPath } from "url";
L23357: var execAsync = promisify(exec);
L23358: function getPackageRoot2() {
High
•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:
(req.headers["authorization"] || req.headers["x-api-key"]) ?? ""
const { baseUrl, authHeaders } = resolveLiteLLMTarget();
const endpoint = `${baseUrl}/v1/embeddings`;
// Optional attribution target for API keys (type "api"): tags requests
Some tools collect credentials (API keys, passwords, tokens) through a secure form shown directly to the user in the chat UI. Credentials are never entered in the conversation itse...
if (output && typeof output === "object" && output.oauth?.authorizationUrl) {
read them with the readFile tool. Files you produce yourself (via writeFile or via shell
this request (readFile, writeFile, bas
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 unpkg•Manifest-reachable source resolves another installed package, overwrites its runtime code, and injects package-defined remote behavior.
dist/index.js:
"Cannot have multiple types in the same job, must be one of the following: embedder, workflow, processor or eval."
description: `This tool calls an agent named: ${instance.name}. The agent does the following: ${instance.description ?? ""}.`,
Example: {url: https://www.google.com, title: Google, snippet: The result of the web search.}
read them with the readFile tool. Files you produce yourself (via writeFile or via shell
// assistant's response, no follow-up text turn is wanted (same
means: read its SKILL.md with the readFile tool, follow the instructions inside it, and run
this request (readFile, writeFile, bash, plus whatever else is regis
High
Entrypoint Foreign Package Code Overwrite
Manifest-reachable source overwrites another installed package with package-defined remote behavior.
dist/index.jsView on unpkg104var redisServer = {
L105: host: process.env.REDIS_HOST ?? "",
L106: port: process.env.REDIS_PORT ?? "",
...
L371: };
L372: var resolveLiteLLMConfigPath = () => process.env.LITELLM_CONFIG_PATH ?? resolve(process.cwd(), "./config.litellm.yaml");
L373: var parseEmbeddingModels = (configPath) => {
...
L529: contentType: type,
L530: metadata: {
L531: ...metadata,
...
L571: const text = this.encoder.decode(tokens);
L572: return new TextDecoder().decode(text);
L573: }
High
Command Output Exfiltration
Source combines command execution, command-output handling, and outbound requests; review data flow before blocking.
dist/index.jsView on unpkg · L104104var redisServer = {
L105: host: process.env.REDIS_HOST ?? "",
L106: port: process.env.REDIS_PORT ?? "",
...
L371: };
L372: var resolveLiteLLMConfigPath = () => process.env.LITELLM_CONFIG_PATH ?? resolve(process.cwd(), "./config.litellm.yaml");
L373: var parseEmbeddingModels = (configPath) => {
...
L529: contentType: type,
L530: metadata: {
L531: ...metadata,
...
L571: const text = this.encoder.decode(tokens);
L572: return new TextDecoder().decode(text);
L573: }
Low
Weak Crypto
Package source references weak cryptographic algorithms.
dist/index.jsView on unpkg · L104ee/python/setup.shView file
•path = ee/python/setup.sh
kind = build_helper
sizeBytes = 10741
magicHex = [redacted]
Medium
Ships Build Helper
Package ships non-JavaScript build or shell helper files.
ee/python/setup.shView on unpkgee/python/transcription/tests/__init__.pyView file
•path = ee/python/transcription/tests/__init__.py
kind = payload_in_excluded_dir
sizeBytes = 0
High
Payload In Excluded Dir
Package hides binary, compressed, or executable-looking payloads in test/fixture/hidden paths.
ee/python/transcription/tests/__init__.pyView on unpkgdist/chunk-7AZH4ETH.jsView file
•stage = ast_semantic_analysis; reason = ast_alias_growth_limit_exceeded; limitedFiles = 1
High
Semantic Analysis Limited
A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.
dist/chunk-7AZH4ETH.jsView on unpkgFindings
1 Critical8 High6 Medium7 Low
CriticalRed Install Lifecycle Scriptpackage.json
HighInstall Time Lifecycle Scriptspackage.json
HighChild Processbin/backend.cjs
HighShelldist/index.js
HighCredential Redirect Persistencedist/index.js
HighEntrypoint Foreign Package Code Overwritedist/index.js
HighCommand Output Exfiltrationdist/index.js
HighPayload In Excluded Diree/python/transcription/tests/__init__.py
HighSemantic Analysis Limiteddist/chunk-7AZH4ETH.js
MediumAmbiguous Install Lifecycle Scriptpackage.json
MediumNetwork
MediumEnvironment Vars
MediumShips Build Helperee/python/setup.sh
MediumStructural Risk Force Deep Review
MediumWildcard Dependency
LowNon Install Lifecycle Scripts
LowScripts Present
LowWeak Cryptodist/index.js
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
LowNo License
Affected versions and remediation
This report applies to @exulu/backend@3.6.0.
See all versions of @exulu/backend for other recorded verdicts.
- Review the evidence and your use of @exulu/backend@3.6.0 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.