registry / @xfive/chisel-ai-toolkit / 1.2.0

@xfive/chisel-ai-toolkit@1.2.0

Chisel AI toolkit — skills, rules and reference docs for AI coding agents working on Chisel WordPress themes

AI Security Review

scanned 6d ago · by lpm-firewall-ai

The postinstall changes project agent configuration, but the inspected source ties that change to the package’s first-party core-directory guard and confines it to a detected Chisel theme. The hook’s scope and cleanup behavior are limited, so no concrete attack surface remains.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

Installing the package runs its postinstall script; the installer acts when it finds a Chisel theme.

Impact

The hook can prevent edits to the theme’s core directory. The inspected behavior does not establish credential collection, external communication, or broad agent control.

Mechanism

The installer copies bundled toolkit files and adds a project-level Claude hook that checks file paths and blocks writes inside core/.

AI rationale

Decision evidence

public snapshot

AI called this Clean at 91.0% confidence as Benign with low false-positive risk.

Evidence for AI clean decision

package.json runs install.js from postinstall, and the installer returns without writing in dry-run mode.
The installer targets a detected Chisel theme and copies the package’s bundled skills, rules, and hook files there.
The installed Claude hook checks the requested file path and blocks writes only inside the theme’s core directory.
The settings helper adds a PreToolUse hook that runs the bundled guard; it does not add unrelated approval grants.
Uninstall removes the toolkit’s managed content and hook while preserving other settings entries.
The package manifest is written under the project’s .claude directory as installation bookkeeping.

Evidence against

Behavioral surface

Source

ChildProcessCryptoEnvironmentVarsFilesystem

Supply chain

HighEntropyStrings

ManifestNo manifest risk signals triggered.

scanned 8 file(s), 32.5 KB of source

Source & flagged code

6 flagged

package.jsonView file

•scripts.postinstall = node install.js

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 1

High

Semantic Analysis Limited

A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.

package.jsonView on unpkg

•scripts.postinstall = node install.js

Medium

Ambiguous Install Lifecycle Script

Install-time lifecycle script is not statically allowlisted and needs review.

package.jsonView on unpkg

install.jsView file

1Install-time AI-agent control hijack evidence: Install-time helper chain writes an AI-agent control surface: L17: const value = () => inline ?? argv[++i]; L18: if (flag === "--agent" || flag === "--agents") options.agents = value().split(","); L19: else if (flag === "--root") options.root = value(); ... L27: const roots = resolveRoots({ override: options.root }); L28: const agents = options.agents || DEFAULT_AGENTS; L29: Referenced control path from install.js: L17: const value = () => inline ?? argv[++i]; L18: if (flag === "--agent" || flag === "--agents") options.agents = value().split(","); L19: else if (flag === "--root") options.root = value(); ... L27: const roots = resolveRoots({ override: options.root }); L28: const agents = options.agents || DEFAULT_AGENTS; L29: Write operation

Critical

Ai Agent Control Hijack

Source creates an unconsented AI-agent control surface through install-time mutation or a default unauthenticated remote skill channel.

install.jsView on unpkg · L1

uninstall.jsView file

•Runtime or CLI code links a write operation to an explicit external AI-agent control path: #!/usr/bin/env node "use strict"; const fs = require("node:fs"); const path = require("node:path"); const pkg = require("./package.json"); const { resolveroots } = require("./lib/roots"); const manifest = require("./lib/manifest"); const { begin, end } = require("./lib/emit"); const settings = require("./lib/settings"); const managed_memory_files = new set(["claude.md", "agents.md"]); function stripmanagedblock(content) { const start = content.indexof(begin); const end = content.indexof(end); if (start === -1 || end === -1 || end < start) return content; return (content.slice(0, start) + content.slice(end + end.length)).replace(/\n{3,}/g, "\n\n").trimstart(); } function pruneemptydirs(di

High

External Ai Agent Control Surface Mutation

Runtime or CLI source writes behavior-bearing configuration into a user or project AI-agent control surface.

uninstall.jsView on unpkg

lib/emit.jsView file

•Manifest-trigger-reachable source links an external AI-agent control path to a behavior-bearing write operation. function render(text, vars) { return text.replace(/\{\{(\w+)\}\}/g, (match, key) => (key in vars ? vars[key] : match)); } const text_extensions = new set([".md", ".mdc", ".json", ".txt", ".yml", ".yaml"]); function copytree(source, target, vars, record) { fs.mkdirsync(target, { recursive: true }); for (const entry of fs.readdirsync(source, { withfiletypes: true })) { const from = path.join(source, entry.name); const to = path.join(target, entry.name); if (entry.isdirectory()) { copytree(from, to, vars, record); } else if (text_extensions.has(path.extname(entry.name))) { fs.writefilesync(to, render(fs.readfilesync(from, "utf8"), vars)); r

High

Trigger Reachable External Ai Agent Control Surface Mutation

Manifest-trigger-reachable source writes behavior-bearing configuration into a user or project AI-agent control surface.

lib/emit.jsView on unpkg

Source & flagged code

6 flagged · loading source

package.jsonView file

•scripts.postinstall = node install.js

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 1

High

Semantic Analysis Limited

A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.

package.jsonView on unpkg

•scripts.postinstall = node install.js

Medium

Ambiguous Install Lifecycle Script

Install-time lifecycle script is not statically allowlisted and needs review.

package.jsonView on unpkg

install.jsView file

Findings

1 Critical4 High3 Medium4 Low

CriticalAi Agent Control Hijackinstall.js

HighInstall Time Lifecycle Scriptspackage.json

HighExternal Ai Agent Control Surface Mutationuninstall.js

HighTrigger Reachable External Ai Agent Control Surface Mutationlib/emit.js

HighSemantic Analysis Limitedpackage.json

MediumAmbiguous Install Lifecycle Scriptpackage.json

MediumEnvironment Vars

MediumStructural Risk Force Deep Review

LowNon Install Lifecycle Scripts

LowScripts Present

LowFilesystem

LowHigh Entropy Strings

Affected versions and remediation

This report applies to @xfive/chisel-ai-toolkit@1.2.0.

See all versions of @xfive/chisel-ai-toolkit for other recorded verdicts.

Review the evidence and your use of @xfive/chisel-ai-toolkit@1.2.0 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: 2026-10-17 04:09:11.227Z (UTC).

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