Caveat verdict

ecommerce-listing

ecommerce-listing-skill

88
๐ŸŸข Trusted
No high-risk patterns surfaced by the deep scan โ€” automated capability review, not behavioral proof.

Encodes data AND uses eval

E-commerce product listing extractor that uses eval of bundled local Python scripts to inject browser DOM commands; the pattern is standard for browser-act skills, scripts are locally installed (not fetched remotely), and behavior matches the stated purpose of reading publicly visible page content.

Automated static analysis โ€” not a human review. Caveat flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.

48
security
80
transparency
70
maintenance

What it does

These are capability combinations: each listed behavior occurs in the skill, but Caveat detects co-occurrence โ€” it does not verify that one flows into another. Read the code to confirm a live chain.

Capability combination critical

Encodes data AND uses eval โ€” the obfuscated-execution pattern (atob + eval; data-flow not verified)

LLM05 ยท ASI05 ยท ASI10

Findings (2)

Pattern match medium

Instructs covert action โ€” may act without user awareness

SKILL.md ยท prose ยท downgraded ยท silently

Pattern match low

References agent memory files

SKILL.md ยท prose ยท downgraded ยท memory.md

Why the tier is capped

Execution sink present in raw bytes (Hard Floor: class B). Final tier capped at Caution โ€” cannot be lifted by any downgrade, example-payload opt-in, or allowlist.

Permissions & capabilities

No declared permissions โ€” minimal attack surface.

dynamic_evaldata_encoding
Check another skill Browse the registry Auditing your own skills or configs? Use the API