What the reviewer found
ClawGuard is a defensive threat-detection skill: every flagged string (the curl exfil to evil.com, ls ~/.ssh/, the regex for hidden HTML-comment instructions) is a worked example of an attack pattern the tool is built to detect and block, shown in its own documentation. It does not perform any of these actions itself.
example-or-fixture— an example or test fixturedual-use— a security tool that can be misused
What was read
The file as it ships in SafeAgent-Beihang/clawguard:
detect-skill/SKILL.md
What the static scan said
The scan flagged 7things. The reviewer kept 0 and dismissed 7 as false.
P1Instruction-override phrasing — false positiveP2Hidden instructions — false positiveE1Sends data to an external URL — false positiveE3Enumerates the file system for secrets — false positivePE2Asks for root — false positivePE3Reaches for credential files — false positiveNETMakes network calls — false positive
How this review was made
Sonnet 5 read the files above on 7 September 2026 and answered three questions: is it dangerous to whoever installs it, is each scanner finding real, and what should the installer know. The verdict is bound to the file's hash; when the file changes, it is scanned afresh and reviewed again. A script that changes while the definition does not is not re-reviewed — that is a known gap. How the scan and the review work.