clawguard
PassAudited by ClawScan on May 1, 2026.
Overview
clawguard appears to be a coherent local security-review tool; users should mainly notice that it scans the chosen folder and saves review reports locally.
This skill looks appropriate for local review of third-party OpenClaw skills. Before installing or using it, remember that it reads files under the path you provide and saves reports locally, so point it only at the intended skill folder and clear saved reports if they contain sensitive snippets.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If the user or agent points the tool at a broad or private directory instead of a specific skill folder, more local files may be inspected than intended.
The scanner walks the entire user-supplied skill path and collects text-like files for review. This is aligned with the stated security-review purpose, but the chosen path controls how much local content is read.
for root, dirnames, filenames in os.walk(skill_path_abs): ... if is_text_file(full_path): files.append(full_path)
Use it only on the specific third-party skill directory you want reviewed, and avoid passing home directories, workspaces, or unrelated private folders.
Review results may remain on disk after the scan, including snippets from files that were flagged during review.
The tool saves scan reports into a persistent local reports.json file under workspace memory. Those reports can include matched evidence snippets and paths from reviewed skill files.
MEMORY_DIR = os.path.join(get_workspace_root(), "memory", "clawguard") REPORTS_PATH = os.path.join(MEMORY_DIR, "reports.json") ... data["reports"][report["report_id"]] = report
Keep the workspace private, review or delete saved reports if they may contain sensitive snippets, and consider adding redaction or retention controls if maintaining this skill.