supply-chain-defense
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The tests/run.sh script generates temporary files containing shell command execution patterns (e.g., curl | sh) and Base64-encoded payloads. These are used as deterministic test fixtures to ensure the skill's detection engine correctly identifies malicious behavior.
- [EXTERNAL_DOWNLOADS]: Several scripts perform network requests to external services: scripts/preinstall-check.sh queries public package registries including npm, PyPI, Packagist, and Crates.io to verify release ages; scripts/phone-home-monitor.ps1 performs RDAP lookups via rdap.org to check domain registration dates; scripts/postinstall-audit.py optionally checks the npm registry to see if a flagged version has been unpublished.
- [DATA_EXFILTRATION]: The scripts/integrity-audit.sh and scripts/postinstall-audit.py scripts are designed to access sensitive configuration paths—such as .aws/credentials, SSH keys, and browser profile data—to audit for unauthorized access or persistent malware hooks. This data is processed locally and not transmitted to external servers.
- [COMMAND_EXECUTION]: The skill invokes various command-line tools for analysis: scripts/integrity-audit.sh uses zizmor for GitHub Actions security analysis; scripts/postinstall-audit.py can invoke guarddog and semgrep for deep behavioral scanning of packages; scripts/config-drift-check.py uses git to analyze staged changes and repository history.
- [PERSISTENCE_MECHANISMS]: The script scripts/phone-home-monitor.ps1 includes functionality to register a Windows Scheduled Task for continuous outbound connection monitoring at logon.
Audit Metadata