Loading npm security reports…

syndes@0.3.0: AI-agent control-surface risk | LPM Firewall

registry / syndes / 0.3.0

syndes@0.3.0

SynDes — a tamper-evident ledger of everything you do in Claude Code, an efficiency score built from it, and a local dashboard that shows you how you actually work. Splits one shared account between the people on it. macOS, Windows and Linux. Zero depende

AI Security Review

scanned 1mo ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. A global npm install automatically rewrites Claude Code's shared hook configuration. The added hooks run on many Claude Code events, collect event payloads, and launch a detached worker.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

npm install -g syndes, unless an environment guard suppresses postinstall.

Impact

The package gains automatic execution in Claude Code sessions and records session event data in the user's Claude configuration area.

Mechanism

Automatic mutation of Claude Code settings to register persistent command hooks.

Policy narrative

Installing this package globally triggers postinstall, which invokes its installer without requiring the user to run syndes install. The installer copies code into the Claude configuration directory and adds command hooks to shared Claude Code settings for numerous events. Each invoked hook reads the event payload, writes it to a local spool, and can start a detached worker. This is an automatic, foreign AI-agent control-surface modification rather than an explicit user-command setup.

AI rationale

The global-only guard reduces scope but does not obtain explicit consent before postinstall rewrites Claude Code's shared settings and installs persistent hooks. This meets the policy for unconsented install-time mutation of a foreign AI-agent control surface.

Evidence

package.jsonbin/postinstall.mjssrc/install.mjssrc/settings.mjsruntime/hook.mjs~/.claude/settings.json~/.claude/settings.json.syndes-backup~/.claude/hooks/syndes

Source & flagged code

5 flagged · loading source

package.jsonView file

•scripts.postinstall = node bin/postinstall.mjs

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg

•scripts.postinstall = node bin/postinstall.mjs

Medium

Ambiguous Install Lifecycle Script

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

package.jsonView on unpkg

runtime/paths.mjsView file

12Install-time code directly mutates a foreign AI-agent control surface: L18: import { fileURLToPath } from 'node:url'; L19: import { mkdirSync } from 'node:fs'; L20: ... L26: /** Claude Code honours CLAUDE_CONFIG_DIR; we must resolve it the same way. */ L27: export const claudeDir = process.env.CLAUDE_CONFIG_DIR?.trim() || join(homedir(), '.claude'); L28: L29: /** True when the config directory has been redirected. An isolated install stays isolated. */ L30: export const isIsolated = Boolean(process.env.CLAUDE_CONFIG_DIR?.trim()); L31: ... L101: export function projectConfigFile(cwd) { L102: return join(cwd, '.claude', 'syndes.json'); L103: } Write operation from src/briefing.mjs: L12: import { isEnabled as sharingOn, teamConfig } from '../sync/index.mjs'; L13: import { existsSync, write

Critical

Ai Agent Control Hijack

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

runtime/paths.mjsView on unpkg · L12

•Manifest-trigger-reachable source links an external AI-agent control path to a behavior-bearing write operation. import { homedir } from 'node:os'; import { dirname, join, resolve } from 'node:path'; import { fileurltopath } from 'node:url'; import { mkdirsync } from 'node:fs'; const runtimedir = dirname(fileurltopath(import.meta.url)); export const packageroot = dirname(runtimedir); export const claudedir = process.env.claude_config_dir?.trim() || join(homedir(), '.claude'); export const isisolated = boolean(process.env.claude_config_dir?.trim()); export const sett

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.

runtime/paths.mjsView on unpkg

collect/git.mjs#virtual:normalized:round1View file

•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.

collect/git.mjs#virtual:normalized:round1View on unpkg