Skillv1.0.1

ClawScan security

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

Scanner verdict

BenignApr 15, 2026, 11:09 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: An instruction-only Hong Kong hiking planning skill whose requirements and instructions are consistent with its stated purpose and only call publicly documented government APIs and included reference files.
Guidance: This skill appears coherent and uses only public, official data sources (Hong Kong Observatory, AFCD/hiking.gov.hk, data.gov.hk). Before installing: (1) note that the package source/homepage is missing — verify the publisher if provenance matters; (2) understand realtime limitations: the skill relies on external APIs for weather, ETA, and facility status and explicitly warns it cannot guarantee minute-level updates — always verify critical, time-sensitive details with official operators before you travel; (3) no credentials are requested, so there is no secret-exfiltration risk from the skill itself; (4) do not rely on it for emergency medical or rescue decisions — follow official emergency contacts (999) and AFCD/HKO notices.
Findings: [no-findings] expected: The regex-based scanner found no matches; this is expected because the skill is instruction-only (no executable code) and bundles plain reference Markdown files and safe API URLs.

Review Dimensions

Purpose & Capability: okThe name/description (HK hiking, transport, weather, facilities, safety) matches the manifest and runtime instructions. The skill requires no binaries, no environment variables, and no installs — all of which are proportionate to a reference/information skill. URLs and data sources referenced (HKO, AFCD/hiking.gov.hk, data.gov.hk) are appropriate for the stated functionality.
Instruction Scope: okSKILL.md limits operations to reading bundled reference files and calling public, documented endpoints (e.g., HKO weather API, HKO GeoJSON, data.gov.hk datasets). Example code snippets show plain GET requests only. The instructions explicitly warn not to fabricate real-time data and to surface limitations, which reduces scope creep. No instructions request unrelated system files, secrets, or network endpoints.
Install Mechanism: okThere is no install spec and no code files to execute — this is instruction-only. That is the lowest-risk install mechanism and consistent with the skill's informational purpose.
Credentials: okThe skill requests no environment variables, credentials, or config paths. All external access is to public government APIs and datasets that do not require credentials, which is proportional to the task.
Persistence & Privilege: okFlags show always:false and normal user-invocable/autonomous settings. The skill does not request persistent or elevated privileges, nor does it instruct modifying other skills or system-wide configs.