registry / AI-agent control-surface risk / @shendu-sdt/sd-zentao-cli
@shendu-sdt/sd-zentao-cli
SD internal ZenTao CLI for bug, task, and workhour workflows.
This report shows the evidence for @shendu-sdt/sd-zentao-cli@0.1.14, the only known version.
AI Security Review
scanned 4mo ago · by lpm-firewall-aiStatic inspection confirms a postinstall hook mutates broad/foreign AI-agent control surfaces without requiring an explicit setup command, which matches the firewall block policy. The rest of the CLI appears package-aligned and does not show credential exfiltration or remote payload execution, but the install-time AI control-surface write is sufficient to block.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
npm install runs package.json postinstall
Impact
Can alter AI assistant behavior on the host without an explicit user setup command; no separate exfiltration or destructive payload was found.
Mechanism
postinstall AI skill/rule installation across agent config directories
Policy narrative
On installation, npm runs scripts/install-ai-skill.mjs automatically. That script creates multiple AI assistant skill directories in the user's home and copies package-controlled SKILL.md into them, plus removes prior zentao-cli skill directories. The skill changes future agent behavior by instructing assistants to use this package's szt command for ZenTao workflows, including bug and workhour actions.
AI rationale
Static inspection confirms a postinstall hook mutates broad/foreign AI-agent control surfaces without requiring an explicit setup command, which matches the firewall block policy. The rest of the CLI appears package-aligned and does not show credential exfiltration or remote payload execution, but the install-time AI control-surface write is sufficient to block.
Evidence
package.jsonscripts/install-ai-skill.mjsskills/sd-zentao-cli/SKILL.mddist/index.jsdist/config.jsdist/api.js~/.codex/skills/sd-zentao-cli/SKILL.md~/.agents/skills/sd-zentao-cli/SKILL.md~/.claude/skills/sd-zentao-cli/SKILL.md~/.cursor/skills/sd-zentao-cli/SKILL.md~/.qoder/skills/sd-zentao-cli/SKILL.md~/.gemini/config/skills/sd-zentao-cli/SKILL.md~/.antigravity/skills/sd-zentao-cli/SKILL.md~/.windsurf/skills/sd-zentao-cli/SKILL.md
Decision evidence
public snapshotAI identified this as a dangerous AI-agent capability at 94.0% confidence with low false-positive risk.
Evidence for policy risk
- package.json defines postinstall: node scripts/install-ai-skill.mjs
- scripts/install-ai-skill.mjs copies SKILL.md into broad AI-agent directories under ~/.codex, ~/.agents, ~/.claude, ~/.cursor, ~/.gemini, etc.
- postinstall also removes old zentao-cli skill directories across the same agent surfaces
- SKILL.md instructs agents to prefer szt and guides bug/workhour write workflows
Evidence against
- No credential exfiltration found; saved ZenTao credentials stay in ~/.szt-cli/config.json with 0600 permissions
- Network use is ZenTao-aligned via configured URL/default https://pm.sdpjw.com/index.php and REST/legacy API calls
- child_process use is limited to invoking install-ai-skill.mjs and retrying the same CLI after relogin; git execFile supports user-invoked work scan
- No eval/vm/dynamic remote payload loading found
Behavioral surface
ChildProcessEnvironmentVarsFilesystemShell
HighEntropyStrings
Source & flagged code
3 flagged · loading sourcepackage.jsonView file
•scripts.postinstall = node scripts/install-ai-skill.mjs
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkgdist/index.jsView file
166patternName = generic_password
severity = medium
line = 166
matchedText = await cl... });
Medium
scripts/install-ai-skill.mjsView file
1Install-time AI-agent control hijack evidence:
L1: #!/usr/bin/env node
L2: import { copyFileSync, existsSync, mkdirSync, rmSync, writeFileSync } from 'node:fs';
L3: import { dirname, join, resolve } from 'node:path';
...
L12: const targets = [
L13: join(homedir(), '.codex', 'skills', skillName, 'SKILL.md'),
L14: join(homedir(), '.agents', 'skills', skillName, 'SKILL.md'),
L15: join(homedir(), '.claude', 'skills', skillName, 'SKILL.md'),
L16: join(homedir(), '.cursor', 'skills', skillName, 'SKILL.md'),
...
L25: const oldTargets = [
L26: join(homedir(), '.codex', 'skills', oldSkillName),
L27: join(homedir(), '.agents', 'skills', oldSkillName),
L28: join(homedir(), '.claude', 'skills', oldSkillName),
Payload evidence from skills/sd-zentao-cli/SKILL.md:
L18: ```bash
L19: npm install -g --regis
Findings
1 Critical1 High3 Medium6 Low
CriticalAi Agent Control Hijackscripts/install-ai-skill.mjs
HighInstall Time Lifecycle Scriptspackage.json
MediumSecret Patterndist/index.js
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowNon Install Lifecycle Scripts
LowScripts Present
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
LowNo License
Affected versions and remediation
This report applies to @shendu-sdt/sd-zentao-cli@0.1.14.
The versions list below shows the verdict for each recorded version.
- Avoid installing @shendu-sdt/sd-zentao-cli@0.1.14. 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: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.