registry / Credential exfiltration / @bysking/b-code / 1.0.2

@bysking/b-code@1.0.2

从零构建的编码助手 — 你自己的 B-Code，运行在终端里

AI Security Review

scanned 9d ago · by lpm-firewall-ai

The active CLI can send an Anthropic API key to the package-selected DeepSeek endpoint when the user retains the default URL. This is concrete credential disclosure, so publishing should be blocked.

Static reason

One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review; source fingerprint signature matched known malicious package; routed for review

Trigger

Run the CLI without an existing settings file, choose Anthropic, enter an Anthropic API key, and retain the default URL.

Impact

The Anthropic API key can be disclosed to the unrelated DeepSeek endpoint.

Mechanism

Setup saves the key and base URL; runtime assigns them to Anthropic configuration and sends authenticated model requests to the configured base URL.

Attack narrative

The CLI’s first-run flow offers Anthropic but supplies DeepSeek as the default base URL. A user who enters an Anthropic key and retains that package-selected URL has the key stored in Anthropic configuration and used for requests to DeepSeek. The URL is displayed and can be changed, but retaining a publisher-provided default does not establish recipient-specific consent to disclose the Anthropic credential.

AI rationale

The active CLI can send an Anthropic API key to the package-selected DeepSeek endpoint when the user retains the default URL. This is concrete credential disclosure, so publishing should be blocked.

Evidence

package.jsondist/cli.mjs~/.b-code/settings.json

Network endpoints1

api.deepseek.com

Decision evidence

public snapshot

AI called this Malicious at 89.0% confidence as Malware with low false-positive risk.

Evidence for block

The manifest exposes dist/cli.mjs as the package CLI, and the CLI starts first-run setup when no settings file exists.
Setup defaults to the OpenAI-compatible provider and DeepSeek URL, prompts for an API key before the URL, and allows the user to choose Anthropic.
The setup saves the chosen key and URL; for Anthropic, runtime configuration maps them to Anthropic credentials and base URL, which the Anthropic client uses for requests.
The CLI also gathers CLAUDE.md and AGENTS.md as project instructions for its model context, a normal coding-assistant feature.

Evidence against

Setup interactively asks for a provider and displays a base URL prompt, so a user can notice or change the recipient.
The URL and model can be customized before setup saves the configuration.

Behavioral surface

Source

ChildProcessCryptoEnvironmentVarsFilesystemNetworkShell

Supply chain

HighEntropyStrings

Source & flagged code

10 flagged

dist/cli.mjsView file

13${A} L14: ${E}`)}return A},Z=N7(N2)});var _u,Hw=ee(()=>{yt();Ft();it();Ot();_u=class extends we{retrieve(e,r={},n){let{betas:i}=r??{};return this._client.get(Z`/v1/deployment_runs/${e}?beta=... L15: `);return e[e.length-1]===""&&e.pop(),e}function IX(t,e){return/^[\x20-\x7E]+$/.test(e)&&!/[\\^#]/.test(e)&&!(t==="unzip"&&e.startsWith("-"))}function QX(t,e,r){let n=hM(e),i=hM(r)...

High

Child Process

Package source references child process execution.

dist/cli.mjsView on unpkg · L13

492Expecting one of '${n.join("', '")}'`);return this._lifeCycleHooks[e]?this._lifeCycleHooks[e].push(r):this._lifeCycleHooks[e]=[r],this}exitOverride(e){return e?this._exitCallback=e... L493: - already used by option '${r.flags}'`)}this._initOptionGroup(e),this.options.push(e)}_registerCommand(e){let r=i=>[i.name()].concat(i.aliases()),n=r(e).find(i=>this._findCommand(... L494: - either make a new Command for each call to parse, or stop storing options as properties`);this._name=this._savedState._name,this._scriptPath=null,this.rawArgs=[],this._optionValu...

High

Shell

Package source references shell execution.

dist/cli.mjsView on unpkg · L492

5`);let n;for(;(n=JK.exec(r))!=null;){let i=n[1],s=n[2]||"";s=s.trim();let A=s[0];s=s.replace(/^(['"`])([\s\S]*)\1$/mg,"$2"),A==='"'&&(s=s.replace(/\\n/g,` L6: `),s=s.replace(/\\r/g,"\r")),e[i]=s}return e}function qK(t){t=t||{};let e=C1(t);t.path=e;let r=$r.configDotenv(t);if(!r.parsed){let A=new Error(`MISSING_DATA: Cannot parse ${e} for... L7: `):[]}};xi=new WeakMap,vi=new WeakMap;ja.NEWLINE_CHARS=new Set([` ... L13: ${A} L14: ${E}`)}return A},Z=N7(N2)});var _u,Hw=ee(()=>{yt();Ft();it();Ot();_u=class extends we{retrieve(e,r={},n){let{betas:i}=r??{};return this._client.get(Z`/v1/deployment_runs/${e}?beta=... L15: `);return e[e.length-1]===""&&e.pop(),e}function IX(t,e){return/^[\x20-\x7E]+$/.test(e)&&!/[\\^#]/.test(e)&&!(t==="unzip"&&e.startsWith("-"))}function QX(t,e,r){let n=hM(e),i=

High

Same File Env Network Execution

A single source file combines environment access, network access, and code or shell execution; review context before blocking.

dist/cli.mjsView on unpkg · L5

•Manifest-reachable source captures an API credential, sends it to a fixed unofficial gateway, and persists that redirection in agent or shell configuration. dist/cli.mjs: `),s=s.replace(/\\r/g,"\r")),e[i]=s}return e}function qK(t){t=t||{};let e=C1(t);t.path=e;let r=$r.configDotenv(t);if(!r.parsed){let A=new Error(`MISSING_DATA: Cannot parse ${e} for... ${E}`)}return A},Z=N7(N2)});var _u,Hw=ee(()=>{yt();Ft();it();Ot();_u=class extends we{retrieve(e,r={},n){let{betas:i}=r??{};return this._client.get(Z`/v1/deployment_runs/${e}?beta=... `);return e[e.length-1]===""&&e.pop(),e}function IX(t,e){return/^[\x20-\x7E]+$/.test(e)&&!/[\\^#]/.test(e)&&!(t==="unzip"&&e.startsWith("-"))}function QX(t,e,r){let n=hM(e),i=hM(r)... `,{flag:"wx",mode:384}),["-xf",e,"-C",r,"-X",i])}async function DX(t,e){let r

High

Credential Redirect Persistence

Manifest-reachable source sends a prompted API credential to a fixed unofficial gateway and persists the redirection.

dist/cli.mjsView on unpkg

3const require = __bcodeCreateRequire(import.meta.url); L4: var FK=Object.create;var lw=Object.defineProperty;var RK=Object.getOwnPropertyDescriptor;var NK=Object.getOwnPropertyNames;var MK=Object.getPrototypeOf,TK=Object.prototype.hasOwnPr... L5: `);let n;for(;(n=JK.exec(r))!=null;){let i=n[1],s=n[2]||"";s=s.trim();let A=s[0];s=s.replace(/^(['"`])([\s\S]*)\1$/mg,"$2"),A==='"'&&(s=s.replace(/\\n/g,` L6: `),s=s.replace(/\\r/g,"\r")),e[i]=s}return e}function qK(t){t=t||{};let e=C1(t);t.path=e;let r=$r.configDotenv(t);if(!r.parsed){let A=new Error(`MISSING_DATA: Cannot parse ${e} for... L7: `):[]}};xi=new WeakMap,vi=new WeakMap;ja.NEWLINE_CHARS=new Set([` ... L13: ${A} L14: ${E}`)}return A},Z=N7(N2)});var _u,Hw=ee(()=>{yt();Ft();it();Ot();_u=class extends we{retrieve(e,r={},n){let{betas:i}

High

Command Output Exfiltration

Source combines command execution, command-output handling, and outbound requests; review data flow before blocking.

dist/cli.mjsView on unpkg · L3

High

Remote Agent Bridge

Source exposes local file and command tools to a remote model endpoint.

dist/cli.mjsView on unpkg · L5

3Trigger-reachable command-output exfiltration chain: manifest.main -> dist/cli.mjs L3: const require = __bcodeCreateRequire(import.meta.url); L4: var FK=Object.create;var lw=Object.defineProperty;var RK=Object.getOwnPropertyDescriptor;var NK=Object.getOwnPropertyNames;var MK=Object.getPrototypeOf,TK=Object.prototype.hasOwnPr... L5: `);let n;for(;(n=JK.exec(r))!=null;){let i=n[1],s=n[2]||"";s=s.trim();let A=s[0];s=s.replace(/^(['"`])([\s\S]*)\1$/mg,"$2"),A==='"'&&(s=s.replace(/\\n/g,` L6: `),s=s.replace(/\\r/g,"\r")),e[i]=s}return e}function qK(t){t=t||{};let e=C1(t);t.path=e;let r=$r.configDotenv(t);if(!r.parsed){let A=new Error(`MISSING_DATA: Cannot parse ${e} for... L7: `):[]}};xi=new WeakMap,vi=new WeakMap;ja.NEWLINE_CHARS=new Set([` ... L13: ${A} L14: ${E}`)}return A},Z=N7(N2)});var _u

High

Trigger Reachable Command Output Exfiltration

A manifest entrypoint or package-local install chain reaches command-output exfiltration behavior.

dist/cli.mjsView on unpkg · L3

•matchType = normalized_sha256 matchedPackage = @bysking/b-code@1.0.1 matchedPath = dist/cli.mjs matchedIdentity = npm:QGJ5c2tpbmcvYi1jb2Rl:1.0.1 similarity = 1.000 summary = normalized source hash matched finalized malicious source

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

dist/cli.mjsView on unpkg

•matchType = malicious_source_fingerprint_signature signature = 22d1d0f7f6501f85 signatureType = suspicious_hashes sourceLabel = final_verdict:malicious matchedPackage = @bysking/b-code@1.0.1 matchedPath = dist/cli.mjs matchedIdentity = npm:QGJ5c2tpbmcvYi1jb2Rl:1.0.1 similarity = 1.000 shingleOverlap = 1 summary = package final verdict is malicious

High

Known Malware Source Fingerprint Signature

Source fingerprint signature matches a known malicious package signature; route for source-aware review.

dist/cli.mjsView on unpkg

Low

Weak Crypto

Package source references weak cryptographic algorithms.

dist/cli.mjsView on unpkg · L3

Source & flagged code

10 flagged · loading source

dist/cli.mjsView file

High

Child Process

Package source references child process execution.

dist/cli.mjsView on unpkg · L13

High

Shell

Package source references shell execution.

dist/cli.mjsView on unpkg · L492

Findings

9 High3 Medium6 Low

HighChild Processdist/cli.mjs

HighShelldist/cli.mjs

HighSame File Env Network Executiondist/cli.mjs

HighCredential Redirect Persistencedist/cli.mjs

HighCommand Output Exfiltrationdist/cli.mjs

HighRemote Agent Bridgedist/cli.mjs

HighTrigger Reachable Command Output Exfiltrationdist/cli.mjs

HighKnown Malware Source Similaritydist/cli.mjs

HighKnown Malware Source Fingerprint Signaturedist/cli.mjs

MediumNetwork

MediumEnvironment Vars

MediumStructural Risk Force Deep Review

LowNon Install Lifecycle Scripts

Low

Affected versions and remediation

This report applies to @bysking/b-code@1.0.2.

See all versions of @bysking/b-code for other recorded verdicts.

Avoid installing @bysking/b-code@1.0.2. Remove it from direct dependencies and check your lockfile for transitive copies.
Choose an independently verified alternative or release. This report does not establish that other versions are safe.
If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.

Evidence last updated: 2026-10-14 06:14:02.890Z (UTC).

Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Affected versions and remediation

Related packages

Other credential exfiltration packages

Behavioral surface

Source & flagged code

Related packages

Other credential exfiltration packages