Back to skill
Skillv1.0.6
ClawScan security
Youtube video scorer · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignFeb 11, 2026, 9:41 AM
- Verdict
- Benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- The skill's files, runtime instructions, and required credential (WORTHCLIP_API_KEY) are consistent with its stated purpose of calling the WorthClip API to score and list YouTube videos.
- Guidance
- This skill is coherent and only needs your WorthClip API key and curl/jq to operate. Before installing: (1) Verify you trust WorthClip and that the base URL (https://greedy-mallard-11.convex.site) is the service deployment you expect (WorthClip documents that domain); (2) Keep your WORTHCLIP_API_KEY secret and only provide keys with appropriate scopes/limits; (3) Note that disable-model-invocation is set, so the model won't call the skill autonomously — you must invoke it; (4) If you ever suspect misuse, revoke/rotate the API key in your WorthClip dashboard.
Review Dimensions
- Purpose & Capability
- okThe name/description require only an API key and the ability to call WorthClip endpoints; the scripts use curl/jq to talk to the documented Convex backend (greedy-mallard-11.convex.site). No unrelated credentials, binaries, or config paths are requested.
- Instruction Scope
- okSKILL.md limits runtime actions to running the provided shell scripts which call WorthClip API endpoints. The scripts only read WORTHCLIP_API_KEY (declared) and do not access other files, system state, or external endpoints beyond the stated base URL.
- Install Mechanism
- okThere is no install spec; this is instruction/script-based and uses existing system tools (curl, jq). Nothing is downloaded or written to disk by an installer.
- Credentials
- okOnly one environment variable is required (WORTHCLIP_API_KEY), which is appropriate for a service-backed API skill. No additional secrets or unrelated variables are requested.
- Persistence & Privilege
- okThe skill is not always-enabled and model invocation is disabled (disable-model-invocation: true), reducing autonomous risk. The skill does not modify other skills or system configs.