Back to skill
Skillv1.0.0
ClawScan security
Report Generator · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
SuspiciousMar 23, 2026, 4:00 AM
- Verdict
- suspicious
- Confidence
- medium
- Model
- gpt-5-mini
- Summary
- The skill's description matches a report generator, but its instructions reference missing scripts and connector docs and do not declare any credentials or configuration required to access POS/ERP or delivery channels — these inconsistencies warrant caution.
- Guidance
- Do not install blindly. Before using, ask the publisher for: (1) the missing data-connectors documentation and the actual scripts (scripts/generate_report.py) that the SKILL.md references; (2) a clear list of required credentials (POS/ERP API keys, WeCom/Telegram tokens) and guidance for storing them securely; (3) explanation of how scheduled delivery is implemented and who can change recipients. If those are not provided, treat the skill as non-functional and avoid giving it access to production credentials or sensitive data. Test in a sandbox with fake data and minimal delivery targets first.
Review Dimensions
- Purpose & Capability
- concernThe skill claims to pull POS/ERP data and deliver reports to channels (WeCom/Telegram). However, it declares no required environment variables, credentials, or config paths. A report generator legitimately needs connectors and auth for POS/ERP and delivery channels; those are not declared or provided.
- Instruction Scope
- concernSKILL.md instructs the agent to run scripts (e.g., scripts/generate_report.py --period ...) and references a data-connectors.md file, but neither the scripts nor data-connectors.md appear in the package. It also refers to a runtime `report_config` (sources, delivery_channel, recipients) without specifying how that config is created or secured. These gaps mean the instructions assume external files/configs that are not present and give the agent broad, unspecified discretion about data sources and recipients.
- Install Mechanism
- okInstruction-only skill with no install spec and no bundled code reduces supply-chain risk. Nothing is downloaded or executed from external URLs by the skill package itself.
- Credentials
- concernNo environment variables or primary credential are declared, yet the skill requires access to POS/APIs and delivery channels which normally need API keys/tokens. The absence of declared secrets is disproportionate to the claimed functionality and leaves unclear where credentials must live and how they're protected.
- Persistence & Privilege
- okalways is false and the skill does not request persistent system changes. It does mention scheduled automatic delivery, but provides no mechanism for establishing schedules or persistent agents; that scheduling behavior is ambiguous but not a declared privilege escalation.