v1.0.1

03 Logistics Alert

ReviewClawScan verdict for this skill. Analyzed May 1, 2026, 11:51 AM.

Analysis

This instruction-only logistics skill has no code, but it should be reviewed because it can push user-provided anomaly judgments into supplier scoring/other skills and presents itself as an official Alibaba internal tool without clear provenance.

GuidanceTreat this skill as advisory unless its official provenance is verified. Do not allow it to automatically update supplier scores, contact suppliers/users, or restrict order intake without human review, verified source data, and an audit trail; limit pasted/uploaded order data to what is necessary.

Findings (5)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

Abnormal behavior control

Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.

Cascading Failures

SeverityMediumConfidenceHighStatusConcern

SKILL.md

如 AE-10 累计 ≥3次/月 → 触发限量评估 ... action_required: [供应商约谈, 暂停新单接入]

Batch anomaly detection can cascade from user-provided order data into supplier reviews and order-intake restrictions without clearly stated verification or approval gates.

User impactA false batch anomaly could escalate into supplier penalties or operational restrictions.

RecommendationRequire independent data verification and explicit human approval before supplier scoring, supplier meetings, order limits, or other operational actions are initiated.

Human-Agent Trust Exploitation

SeverityMediumConfidenceMediumStatusConcern

SKILL.md

author: 阿里巴巴内控部（思维训练版） ... 你是阿里巴巴内控部的物流异常预警专家

The skill frames itself as an Alibaba internal-control expert; with the provided registry source listed as unknown and no homepage, that authority claim is not substantiated by the artifacts.

User impactUsers may over-trust compensation rules or supplier-risk recommendations as official Alibaba guidance.

RecommendationClearly label the skill as unofficial/advisory unless provenance is verified, and require users to confirm decisions against official platform rules.

Agentic Supply Chain Vulnerabilities

SeverityLowConfidenceHighStatusNote

SKILL.md

reference: docs/GLOSSARY.md ... docs/INSUFFICIENCY-HANDLING.md ... docs/RULE-UPDATE-SOP.md ... docs/LINKING-SOP.md

The skill relies on referenced rule, insufficiency-handling, update, and linking documents that are not present in the provided file manifest.

User impactSome rule boundaries and update/linking procedures cannot be checked from the supplied artifacts.

RecommendationBundle the referenced documents or remove/replace those references so users can verify the rules the skill applies.

Sensitive data protection

Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.

Insecure Inter-Agent Communication

SeverityMediumConfidenceHighStatusConcern

SKILL.md

触发 Skill 2 联动：通知 Skill 2：该供应商有 AE-10 事件；Skill 2 追加违规记录（A4维度扣分）

The skill directs another skill to receive supplier incident data and change a scoring record, but it does not define the other skill's identity, permissions, data fields, or approval boundary.

User impactIncorrect or unverified logistics data could be passed to another agent/skill and affect a supplier's internal score.

RecommendationMake cross-skill handoffs opt-in, name the exact receiving skill, specify allowed data fields, and require human confirmation before any score or record is changed.

Memory and Context Poisoning

SeverityMediumConfidenceHighStatusConcern

references/continuation.md

记录每次异常，异常数据纳入供应商风险评分（联动风险评分Skill）

The artifact instructs persistent recording of anomaly data into supplier risk scoring, but does not define validation, retention, correction, or deletion controls.

User impactA mistaken, incomplete, or maliciously supplied order report could persist and influence future supplier decisions.

RecommendationKeep outputs read-only by default; if records are stored, require verified data sources, minimal fields, audit logs, retention limits, and a correction/rollback process.