ClawAudit verdict
clawguard-shield
A prompt injection detection skill that analyzes inputs for encoding patterns, role hijacking, and jailbreak attempts; the data_encoding capability is used to detect encoded attacks (not to exfiltrate), and all behavior is local analysis with no outbound transmission.
⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.
Automated static analysis — not a human review. ClawAudit flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.
Findings (6)
Unicode homoglyph detected — uses lookalike characters to evade pattern matching
SKILL.md · prose
Prompt injection — tries to override agent instructions
SKILL.md · code · ignore previous instructions
Base64 decode (atob) — may hide malicious payloads
SKILL.md · prose · downgraded · atob(
Possible prompt injection — attempts to redefine agent identity
SKILL.md · prose · downgraded · you are now
Uses exec() — may execute shell commands
src/shield.js · prose · downgraded · exec(
String.fromCharCode — can build strings to evade detection
src/shield.js · prose · downgraded · String.fromCharCode
Permissions & capabilities
No declared permissions — minimal attack surface.
data_encoding Is this flag fair?
Thanks — recorded.