registry / @leanandmean/coding-agent / 0.74.1-scramjet.82
@leanandmean/coding-agent@0.74.1-scramjet.82
Coding agent CLI with read, bash, edit, write tools and session management
AI Security Review
scanned 7d ago · by lpm-firewall-aiSource inspection found no install hook or concrete attack behavior. The proxy transfer uses a caller-supplied token and destination, while the flagged similarity findings resolve to ordinary agent functions and example code.
Static reason
One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review
Trigger
No malicious trigger was identified; the inspected runtime functions are invoked through the agent’s normal APIs or examples.
Impact
No confirmed security impact was identified.
Mechanism
No covert collection, unrelated credential transfer, persistence, or destructive behavior was established.
AI rationale
Source inspection found no install hook or concrete attack behavior. The proxy transfer uses a caller-supplied token and destination, while the flagged similarity findings resolve to ordinary agent functions and example code.
Decision evidence
public snapshotAI called this Clean at 91.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- package.json points to the dist entry and lists development, build, and asset-copy scripts, with no install lifecycle hook.
- The flagged shell and process code implements the agent’s local command execution and package-management functions; it does not establish an unrequested command or destructive action.
- The flagged clipboard, RPC, notification, config, and tool-manager code matches their respective product functions; the inspected examples do not establish covert collection or exfiltration.
- The proxy code sends its supplied auth token to its supplied proxy URL as part of an explicitly configured stream function; no package-selected default recipient or credential forwarding caller was found.
- The three inspected first-party runtime dependencies have no install lifecycle hooks in their manifests; their entrypoints export agent, AI, and TUI APIs.
- The flagged DOOM bundle is a generated module loaded by the example DOOM engine; its similarity label alone does not show attack behavior.
- AgentSession manages agent and session state; login guidance points users to provider OAuth or API-key documentation, bash execution accepts a supplied command, and the default thinking level is medium.
Source & flagged code
19 flaggeddist/core/exec.jsView file
3*/
L4: import { spawn } from "node:child_process";
L5: import { StringDecoder } from "node:string_decoder";
High
dist/utils/clipboard.jsView file
42[
L43: "powershell.exe",
L44: [
High
•matchType = normalized_sha256
matchedPackage = @leanandmean/coding-agent@0.74.1-scramjet.81
matchedPath = dist/utils/clipboard.js
matchedIdentity = npm:[redacted]:0.74.1-scramjet.81
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High
Findings
16 High5 Medium5 Low
HighChild Processdist/core/exec.js
HighShelldist/utils/clipboard.js
HighCommand Output Exfiltrationexamples/extensions/doom-overlay/doom/build/doom.js
HighSemantic Analysis Limiteddist/utils/shell.js#virtual:normalized:round1
HighKnown Malware Source Similaritydist/utils/shell.js
HighKnown Malware Source Similarityexamples/extensions/doom-overlay/doom/build/doom.js
HighKnown Malware Source Similaritydist/config.js
HighKnown Malware Source Similaritydist/core/package-manager.js
HighKnown Malware Source Similaritydist/core/tools/bash.js
HighKnown Malware Source Similaritydist/core/resolve-config-value.js
High
Affected versions and remediation
This report applies to @leanandmean/coding-agent@0.74.1-scramjet.82.
See all versions of @leanandmean/coding-agent for other recorded verdicts.
- Review the evidence and your use of @leanandmean/coding-agent@0.74.1-scramjet.82 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.