ClawAudit verdict
skillboss
skillboss-3
A multi-AI gateway skill providing an API wrapper for building full-stack applications with 50+ AI APIs; all network calls go to documented provider endpoints for their stated purposes.
⚠ 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 (11)
Unicode homoglyph detected — uses lookalike characters to evade pattern matching
extensions/remotion/rules/maps.md · code
Attempts to modify skills or system prompts
scripts/commands/document.js · prose · downgraded · edit instructions
Accesses shell history/config
SKILL.md · prose · downgraded · ~/.zshrc
Blob URL — may embed executable content
extensions/remotion/rules/can-decode.md · code · blob:
Accesses sensitive environment variables
extensions/remotion/rules/maps.md · code · process.env.REMOTION_MAPBOX_TOKEN
Downloads executables from external URLs
install/update.sh · prose · downgraded · download from https://
References child_process — can spawn system processes
scripts/api-hub.js · prose · downgraded · child_process
Instructs covert action — may act without user awareness
scripts/lib/client.js · prose · downgraded · Silently
Dynamic import() — loads module at runtime
templates/worker-ecommerce/tailwind.config.js · prose · downgraded · import('
Makes HTTP request to external URL
extensions/remotion/rules/compositions.md · code · fetch(`https://
References agent configuration files
workflows.md · prose · downgraded · CLAUDE.md
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class B/D). Final tier capped at Caution — cannot be lifted by any downgrade, example-payload opt-in, or allowlist.
Permissions & capabilities
No declared permissions — minimal attack surface.
Is this flag fair?
Thanks — recorded.