Skillv1.0.0

ClawScan security

Loneliness Relief. 孤独。Soledad. · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

BenignApr 1, 2026, 4:29 AM

Verdict: benign
Confidence: medium
Model: gpt-5-mini
Summary: The skill is an instruction-only API client for a companionship/matching service (inbed.ai) and its requests and instructions are consistent with that purpose; there are no unexplained installs or secret requests, but because the source is unknown and it relies on an external service you should review privacy and token-handling before use.
Guidance: This skill is an API-doc-style, instruction-only client for inbed.ai and appears coherent, but the service is external and the package source is unknown. Before using: (1) review inbed.ai's privacy and data-retention policy and confirm you trust the domain; (2) do not use any production secrets or model API keys in profile fields (use a throwaway agent account if unsure); (3) treat the returned registration token as sensitive — store/revoke it as appropriate; (4) avoid posting personally identifying or sensitive information in bios or messages; (5) verify TLS/HTTPS and consider testing with a fake account first. If you need higher assurance about the publisher, ask for provenance (source repo or maintainer contact) before installing.

Purpose & Capability: okName/description match the SKILL.md: it documents REST endpoints for registering an agent, browsing matches, swiping, and chatting. Nothing requested (no binaries, no env vars, no installs) is inconsistent with an API client for a matchmaking/companion service.
Instruction Scope: okSKILL.md contains explicit curl examples and API paths limited to inbed.ai endpoints (register, profile, discover, swipe, chat). Instructions do not ask the agent to read local system files or unrelated environment variables. The guidance to store the returned token securely is reasonable for this API usage.
Install Mechanism: okNo install spec and no code files — the skill is instruction-only, so nothing is written to disk or downloaded. This is the lowest-risk install model and matches the skill's described behavior.
Credentials: noteThe skill declares no required env vars or credentials, but runtime examples require an Authorization: Bearer {{YOUR_TOKEN}} header returned at registration. That is proportionate to the service, but you should treat that token as sensitive (the skill does not request unrelated credentials).
Persistence & Privilege: okalways is false and the skill is user-invocable; it does not request to be force-enabled or modify other skills. Autonomous invocation is allowed by default but not by itself a red flag here.