All issues

Search results

HOMOGLYPH_ATTACK HIGH on trailing comments: non-ASCII comment prose misread as mixed-script obfuscation

Status: Open.

#165 In cisco-ai-defense/skill-scanner;

· serkud opened on Aug 25, 2026

[BUG] .git file skip logic contradicts documented "directory only" intent (v2.0.13) (malicious skill example that bypasses this scanner)

bugSomething isn't workingSomething isn't working

Status: Open.

#163 In cisco-ai-defense/skill-scanner;

· nedlir opened on Aug 11, 2026

[BUG] BytecodeAnalyzer fails open without decompyle3/uncompyle6 (malicious skill bypasses this scanner)

bugSomething isn't workingSomething isn't working

Status: Open.

#162 In cisco-ai-defense/skill-scanner;

· nedlir opened on Aug 10, 2026

Behavioral analyzer: 6 false-positive patterns (eval/subprocess name-prefix match, bounded loops, markdown docs, credential-denylist guard, targeted env lookups, comment text)

Status: Open.

#160 In cisco-ai-defense/skill-scanner;

· davinbaker-agent opened on Aug 9, 2026

Pipeline analyzer boundary for remote fetch followed by PowerShell execution

Status: Open.

#150 In cisco-ai-defense/skill-scanner;

· yuehua1213 opened on Aug 3, 2026

[FEATURE] Add skill context to log messages when scanning multiple skills

enhancementNew feature or requestNew feature or request

Status: Open.

#149 In cisco-ai-defense/skill-scanner;

· ojacques opened on Jul 30, 2026