Skillv1.0.0

ClawScan security

Stock Evaluator · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignFeb 11, 2026, 8:32 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's instructions, required assets, and lack of install/credentials are consistent with a self-contained stock analysis tool that performs web searches and produces a detailed report/dashboard — nothing requested appears disproportionate or unrelated to its stated purpose.
Guidance: This skill is internally consistent: it asks the agent to perform web searches and to use public filings (SEC, exchange data, Yahoo/Google Finance) rather than stored secrets or installed binaries. Before installing, confirm that (1) you are comfortable with the agent performing live web lookups and accessing any agent/project memory (it will check whether you already hold the stock), (2) you understand the output is investment analysis (not executed trades) and should not be relied on as sole financial advice, and (3) if you expect a runnable React dashboard you may need to permit the agent to produce code/artifacts or to supply your own hosting/build process because the skill ships only markdown guides, not a packaged web app. If you want to limit exposure, ensure browsing access and memory access are enabled only with your consent.

Review Dimensions

Purpose & Capability: noteName/description match the runtime instructions: the SKILL.md sets out a comprehensive, multi-method stock analysis and a 60+ metric React dashboard. The skill requests no external credentials or installs, which is proportionate. One minor note: the dashboard requirement implies producing a React artifact (code/assets) even though the package contains only guides and no ready-made code; this is plausible (the agent can output code or a JSON dashboard) but the deliverable expectation is heavier than typical instruction-only skills.
Instruction Scope: okSKILL.md narrowly confines actions to stock research: mandated web searches, reading company filings (SEC EDGAR), exchange/financial-data sources, and constructing calculated metrics. It explicitly forbids fabricating numeric data. The only cross-boundary action is a directive to check 'project knowledge' for an existing portfolio (i.e., agent memory/context) — appropriate for investment analysis but worth noting because it will read user/agent context if available.
Install Mechanism: okThere is no install spec and no code to run; the skill is instruction-only. That is low risk and consistent with the described functionality.
Credentials: okThe skill requires no environment variables, credentials, or config paths. Its data needs are satisfied via public web searches and filings (as described), so requested privileges are minimal and appropriate.
Persistence & Privilege: okalways is false and the skill does not request persistent system-level changes or access to other skills' configs. The only notable privilege is that instructions ask the agent to consult project knowledge (agent memory) — this is not a system privilege but may access stored user portfolio/context.