skills/adobe/skills/slicc-handoff/Gen Agent Trust Hub

slicc-handoff

Pass

Audited by Gen Agent Trust Hub on Jul 7, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to fetch protocol and API definitions from sliccy.ai/llms.txt. This is used to ensure the agent uses the correct data structures for the handoff protocol.
  • [COMMAND_EXECUTION]: The provided script scripts/slicc-handoff uses node:child_process to trigger the system's default browser opener (such as open on macOS, xdg-open on Linux, or start on Windows). The script safely encodes user-provided instructions using encodeURIComponent before passing them to the shell, preventing command injection.
  • [DATA_EXFILTRATION]: The skill transmits task context and instructions to localhost:5710 (a local SLICC instance) and to sliccy.ai via URL parameters. This behavior is the core purpose of the skill—handing off work to the browser agent— and the instructions explicitly warn the user not to include sensitive credentials in the handoff text.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 7, 2026, 07:07 AM
Security Audit — agent-trust-hub — slicc-handoff