skills/airtap-ai/airtap-skill/airtap/Gen Agent Trust Hub

airtap

Pass

Audited by Gen Agent Trust Hub on Apr 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONCREDENTIALS_UNSAFE
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/airtap.py executes an external binary named openclaw using subprocess.run to relay task updates.
  • The implementation uses a list of arguments for subprocess.run, which avoids shell interpretation and protects against command injection into the shell itself.
  • The path to the binary can be specified via the --openclaw-bin flag, though the script includes a validation check to ensure the target filename is exactly openclaw.
  • [DATA_EXFILTRATION]: The skill facilitates the transfer of data from the Airtap API to an external destination via the OpenClaw relay system.
  • This behavior is an intended feature of the skill, allowing users to mirror task progress in other communication channels.
  • [CREDENTIALS_UNSAFE]: The skill manages a personal access token by writing it to a local .env file within the skill's directory.
  • Evidence: scripts/airtap_common.py contains the write_token_to_dotenv function which uses set_key to store the AIRTAP_PERSONAL_ACCESS_TOKEN.
  • This is a common pattern for local CLI tools, though it relies on the host environment's file system security to protect the token.
  • [EXTERNAL_DOWNLOADS]: The skill relies on standard third-party libraries for its operation.
  • Evidence: requirements.txt lists requests and python-dotenv as dependencies.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 15, 2026, 03:31 AM