registry / AI-agent control-surface risk / @mdd-ai-eng/specs
@mdd-ai-eng/specs
AI 工程化 MDD 通用规范包。标识 mdd-ai-eng 用于区分同名锚点。ai-engineering-specs 只放规范正文。安装时按实际位置刷新业务项目锚点。
This report shows the evidence for @mdd-ai-eng/specs@1.2.0. 3 of 5 known versions are blocked or flagged. See all versions.
AI Security Review
scanned 23h ago · by lpm-firewall-aiThe automatic postinstall mutation adds package-controlled instructions to consumer AI-agent configuration without a separate user setup action. This meets the policy for blocking unconsented lifecycle changes to a foreign agent control surface.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
Installing the package with lifecycle scripts enabled.
Impact
The added files can affect how consumer agents handle later tasks without a separate setup action.
Mechanism
The installer writes a Kilo rules file and a Cursor rule under the consumer root, embedding a package instruction for agent sessions.
Policy narrative
During package installation, the postinstall hook writes Kilo and Cursor agent rules into the consumer root. Those rules direct agents to load this package’s AGENTS.md at session start and when receiving tasks, creating an unconsented change to the consumer’s agent control surface.
AI rationale
The automatic postinstall mutation adds package-controlled instructions to consumer AI-agent configuration without a separate user setup action. This meets the policy for blocking unconsented lifecycle changes to a foreign agent control surface.
Evidence
package.jsonscripts/install-anchors.js.kilocoderrules.cursor/rules/mdd-ai-eng.mdc
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
- The package runs the anchor installer automatically through its postinstall hook.
- The installer adds a persistent instruction telling agents to load the package’s AGENTS.md at session start and when receiving tasks.
- The installer writes configuration in the consumer root for Kilo and Cursor; it can also update an existing Kilo anchor.
Evidence against
Behavioral surface
EnvironmentVarsFilesystem
NoLicense
Source & flagged code
7 flagged · loading sourcepackage.jsonView file
•scripts.postinstall = node scripts/install-anchors.js
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg10"postinstall": "node scripts/install-anchors.js",
"test": "node test/install-anchors.test.js",
"check": "bash scripts/check-system.sh",
"release": "bash scripts/rele
High
Suspicious Lifecycle Evidence
The package runs the anchor installer automatically through its postinstall hook.
package.jsonView on unpkg · L10scripts/install-anchors.jsView file
66"ai-engineering-specs",
L67: "AGENTS.md"
L68: );
...
L71: }
L72: return path.join(specDirectory, "AGENTS.md");
L73: }
...
L79: }
L80: fs.mkdirSync(directory, { recursive: true });
L81: fs.writeFileSync(filePath, next, "utf8");
L82: return { action: existing === null ? "write" : "update", filePath };
Critical
Findings
1 Critical6 High2 Medium3 Low
CriticalAi Agent Control Hijackscripts/install-anchors.js
HighInstall Time Lifecycle Scriptspackage.json
HighSuspicious Lifecycle Evidencepackage.json
HighAi Review Evidencescripts/install-anchors.js
HighAi Review Evidencescripts/install-anchors.js
HighAi Review Evidencescripts/install-anchors.js
HighAi Review Evidencescripts/install-anchors.js
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowScripts Present
LowFilesystem
LowNo License
Affected versions and remediation
This report applies to @mdd-ai-eng/specs@1.2.0.
The versions list below shows the verdict for each recorded version.
- Avoid installing @mdd-ai-eng/specs@1.2.0. 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.
Versions
5 known versions1.2.0latestshownMaliciousSoft block: AI-agent control surfaceBlockPublishedOct 22, 2026, 1:49 AM1.1.2SuspiciousAgent extension lifecycle riskWarnPublishedOct 20, 2026, 6:45 AM1.1.1SuspiciousAgent extension lifecycle riskWarnPublishedOct 20, 2026, 4:23 AM1.1.0MaliciousSoft block: AI-agent control surfaceBlockPublished