Back to skill
Skillv3.1.1

ClawScan security

波街(Bot Street) · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignApr 23, 2026, 3:24 AM
Verdict
Benign
Confidence
high
Model
gpt-5-mini
Summary
The skill's requests and runtime instructions align with its description as a Bot/marketplace API integration; it does not ask for unrelated credentials, install artifacts, or local file access.
Guidance
This SKILL.md is internally consistent with a Bot marketplace integration. Before installing: (1) only provide agentId/agentKey to trusted platforms — these let the skill act as your Bot; (2) be aware the skill can send private messages and accept tasks on the platform, so restrict autonomous invocation if you don't want the agent to contact others or transact without approval; (3) note the doc instructs callers to re-fetch /skill.md for updates — remote docs can change behavior over time, so re-check permissions when the platform updates; (4) test in a sandbox account first and rotate/revoke agent keys if you stop using the skill.

Review Dimensions

Purpose & Capability
okThe name/description (Bot Street marketplace) match the SKILL.md content: REST API endpoints, agent registration, posting, messaging, task flows and talent marketplace. The skill does not request unrelated binaries, environment variables, or config paths.
Instruction Scope
noteAll instructions are API-focused (relative URLs under current domain, use of x-agent-id/x-agent-key headers, SSE/long-polling for messages). The document does not instruct the agent to read local files or unrelated environment variables. Note: the doc explicitly tells callers to re-fetch /skill.md for updates — this means server-side docs can change expected behavior over time, so runtime behavior may vary if the platform's hosted docs change.
Install Mechanism
okNo install spec and no code files are present (instruction-only), so nothing is downloaded or written to disk at install time.
Credentials
okThe skill declares no required environment variables or primary credential. At runtime the platform expects agentId/agentKey (obtained through user action) — these are proportional to a bot-integration skill and are clearly documented as headers.
Persistence & Privilege
okalways is false and the skill does not request elevated platform-wide persistence. The skill enables typical agent actions (read posts, send DMs, accept tasks) which are expected for a marketplace bot; be aware autonomous invocation combined with messaging capabilities can perform external-facing actions on behalf of the agent if allowed.