registry / @aswless_854771076/ai_short_studio_cli / 0.1.61
@aswless_854771076/ai_short_studio_cli@0.1.61
VVICAT 无限画布项目命令行工具
AI Security Review
scanned 11d ago · by lpm-firewall-aiThe postinstall hook performs first-party AI agent skill setup in the user's Codex skills directory, which warrants a warning under the agent integration policy. Inspected OAuth and bootstrap code use the authentication flow and service endpoint as expected; no concrete attack behavior was established.
Static reason
High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review; previous stored version diff introduced dangerous source
Trigger
npm installation, unless VVICAT_SKIP_SKILL_INSTALL=1 or INIT_CWD identifies the package itself.
Impact
The agent may load instructions for operating the vendor's remote CLI. The inspected source does not establish credential theft or another active attack.
Mechanism
The postinstall hook imports the skill manager, which copies bundled skill files to ~/.codex/skills by default; an environment variable can redirect the destination.
AI rationale
The postinstall hook performs first-party AI agent skill setup in the user's Codex skills directory, which warrants a warning under the agent integration policy. Inspected OAuth and bootstrap code use the authentication flow and service endpoint as expected; no concrete attack behavior was established.
Evidence
package.jsonscripts/postinstall.mjsdist/skill/skill-manager.jsskills/using-vvicat-ai-short-studio-cli/SKILL.md~/.codex/skills/using-vvicat-ai-short-studio-cli
Network endpoints2
ai-short-studio.vvicat.dev127.0.0.1
Decision evidence
public snapshotAI called this Suspicious at 91.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- package.json runs scripts/postinstall.mjs during installation.
- The postinstall skips only on an explicit environment flag or when INIT_CWD resolves to the package itself; otherwise it installs bundled skills.
- The default install target is the user's ~/.codex/skills directory, and installation copies bundled skill content there.
- The installed first-party skill guides an AI agent to operate the vendor's remote CLI; this is an agent extension setup capability.
- Google login uses an OAuth authorization flow and a loopback callback; preflight requests the configured service's bootstrap endpoint.
Evidence against
- Skill installation skips an existing unmanaged or modified destination unless managed state permits updating; the documented command also requires explicit invocation.
- The inspected OAuth flow returns the session from the auth client's code exchange, and preflight uses the selected service base URL.
- No evidence inspected of credential exfiltration, destructive behavior, or remote payload execution.
Behavioral surface
Source & flagged code
6 flagged · loading sourcepackage.jsonView file
•scripts.postinstall = node scripts/postinstall.mjs
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg•scripts.postinstall = node scripts/postinstall.mjs
Medium
Ambiguous Install Lifecycle Script
Install-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgdist/skill/skill-manager.jsView file
1Install-time code directly mutates a foreign AI-agent control surface:
L12: export function defaultSkillsRoot(target) {
L13: return join(homedir(), target === 'codex' ? '.codex' : '.agents', 'skills');
L14: }
...
L41: const backup = `${target}.backup-${process.pid}`;
L42: await mkdir(dirname(target), { recursive: true });
L43: await rm(temporary, { recursive: true, force: true });
L44: await cp(bundledSkillPath(name), temporary, { recursive: true });
L45: await writeFile(join(temporary, MANIFEST), `${JSON.stringify({ packageVersion, checksum }, null, 2)}\n`, 'utf8');
Write operation from dist/skill/sk[redacted]:
L12: export function defaultSkillsRoot(target) {
L13: return join(homedir(), target === 'codex' ? '.codex' : '.agents', 'skills');
L14: }
...
L41: const backup = `${target}.backup-
Findings
1 Critical4 High4 Medium4 Low
CriticalAi Agent Control Hijackdist/skill/skill-manager.js
HighInstall Time Lifecycle Scriptspackage.json
HighTrigger Reachable External Ai Agent Control Surface Mutationdist/skill/skill-manager.js
HighKnown Malware Source Similaritydist/auth/google-oauth.js
HighPrevious Version Dangerous Deltadist/preflight.js
MediumAmbiguous Install Lifecycle Scriptpackage.json
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowScripts Present
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
Affected versions and remediation
This report applies to @aswless_854771076/ai_short_studio_cli@0.1.61.
See all versions of @aswless_854771076/ai_short_studio_cli for other recorded verdicts.
- Review the evidence and your use of @aswless_854771076/ai_short_studio_cli@0.1.61 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.