Back to skill
Skillv0.1.0
ClawScan security
Paper Scout · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMar 7, 2026, 6:27 PM
- Verdict
- benign
- Confidence
- medium
- Model
- gpt-5-mini
- Summary
- The skill's instructions and requirements are broadly consistent with its stated purpose (searching and summarizing robotics papers); it makes no disproportionate environment or credential requests, but its Google Scholar scraping step is vague about tooling and may require browser automation and network activity that you should confirm before installing.
- Guidance
- This skill is coherent for its stated task, but before installing: 1) confirm how the agent will perform the Google Scholar step — what browser automation/tooling it will use and whether it will access your local browser/profile or require credentials; 2) expect the agent to write Markdown files to your Desktop (~/Desktop/YYYY-MM-DD-academic-digest.md); 3) understand that scraping Google Scholar can trigger rate limits, CAPTCHAs or violate terms of service — ask whether the implementation uses polite rate-limiting or official APIs where possible; and 4) if you need stronger guarantees, request the author to specify required tooling and to allow configuring the output path and scraping behavior (or to rely only on CrossRef/API-based sources). If you cannot confirm those points, proceed cautiously.
Review Dimensions
- Purpose & Capability
- okName and description match the instructions: searching CrossRef and (as a fallback) Google Scholar to build Markdown digests. No unrelated credentials, binaries, or config paths are requested, so the requested capabilities line up with the stated purpose.
- Instruction Scope
- noteSKILL.md instructs the agent to query CrossRef, deduplicate/filter results, and save Markdown files to ~/Desktop. It also calls for Google Scholar scraping using 'real-time browser automation' for JS-heavy pages; however it doesn't specify what automation tooling (e.g., puppeteer, selenium) or whether it will use the user's browser/profile. Writing outputs to the user's Desktop is explicit and should be acceptable if expected.
- Install Mechanism
- okNo install spec or code files are included (instruction-only). This is low-risk from an install perspective because nothing is written to disk by an installer. Any runtime downloading or automation tooling would be the agent's responsibility, not specified here.
- Credentials
- okThe skill declares no required environment variables, credentials, or config paths. That is proportionate to a read-only literature-scouting function. Note: Google Scholar scraping sometimes needs proxies, cookies, or logins in practice; the SKILL.md does not request or justify such secrets.
- Persistence & Privilege
- okalways is false and user-invocable is true (normal defaults). The SKILL.md does not request permanent agent-level privileges or changes to other skills or system-wide settings.