Back to skill
Skillv1.0.0
ClawScan security
US Stock Radar · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMar 9, 2026, 2:39 PM
- Verdict
- Benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- The skill's code, instructions, and requirements are consistent with a read-only US stock screening tool that queries public market endpoints and does not request credentials or persist privileged state.
- Guidance
- This skill appears coherent and read-only: it fetches public market data from Yahoo and Stooq, computes heuristic scores, and prints JSON. Before installing/use: 1) verify the script path in SKILL.md matches where you will run it; 2) ensure your environment has Python 3 and the requests library; 3) be aware it makes outbound HTTP requests (so run it in a network environment you trust); and 4) run the quick-audit commands (--sources, --mode screener --json) to confirm outputs, availability fields, and that no unexpected endpoints are contacted. The tool is for signal triage only — treat outputs as heuristic guidance, not trading advice.
Review Dimensions
- Purpose & Capability
- okName/description (US stock radar) match the included script and SKILL.md: the script queries Yahoo Finance and Stooq, computes heuristic metrics, and emits JSON scores/grades. No credentials, unrelated binaries, or cloud permissions are requested. Minor packaging inconsistencies: SKILL.md examples reference a path like skills/us-stock-radar/scripts/us_stock_radar.py while the manifest shows scripts/us_stock_radar.py at repo root; script version string (0.1.1) differs from skill registry version (1.0.0). These are packaging/metadata issues but do not contradict capability.
- Instruction Scope
- okRuntime instructions are narrowly scoped: run the provided Python script in one of three modes and read its JSON output. SKILL.md explicitly enforces read-only behavior (no auth, no trading, no file writes, no outbound messages). The script's network calls are limited to public Yahoo and Stooq endpoints; it does not read arbitrary local files, environment variables, or system configs. Note: the SKILL.md path examples differ from the manifest path — verify where the script will actually be run in your environment.
- Install Mechanism
- okNo install spec is provided (instruction-only with an included script). This minimizes install-time risk; the only runtime requirement is the requests library and Python 3. The script does not attempt to download or execute external installers.
- Credentials
- okThe skill declares no required environment variables, credentials, or config paths and the code does not access os.environ. It only issues HTTP GETs to public data endpoints (Yahoo, Stooq). The absence of secrets or unrelated env access is proportionate to the stated purpose.
- Persistence & Privilege
- okThe skill is not always-enabled and is user-invocable. It does not write files as part of normal operation, does not modify other skills or global settings, and does not request persistent privileges. Note that agent autonomous invocation is allowed by default (disable-model-invocation: false) — this is the platform default and not a red flag here given the read-only nature of the code.