pilot-dns

Pass

Audited by Gen Agent Trust Hub on Jun 21, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the 'pilotctl' command-line tool to register hostnames, find peers, and manage connections. This functionality is essential for its stated purpose of managing a human-friendly naming system for the Pilot Protocol.
  • [EXTERNAL_DOWNLOADS]: Documentation and installation links point to 'pilotprotocol.network' and 'clawhub.ai'. These represent the official project infrastructure and distribution platforms for the protocol and its skills.
  • [PROMPT_INJECTION]: The skill ingests peer hostnames from the network, which creates a surface for indirect prompt injection if a peer provides a crafted hostname string.
  • Ingestion points: Hostname and metadata fields returned by 'pilotctl --json peers' and 'pilotctl --json find' in SKILL.md.
  • Boundary markers: No specific delimiters are implemented in the prompt instructions to isolate untrusted peer data.
  • Capability inventory: The skill uses the Bash tool to execute 'pilotctl' commands, including network resolution and connection initialization.
  • Sanitization: The provided workflow examples utilize 'jq' to extract specific fields from JSON responses, which provides a layer of structural validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 21, 2026, 10:59 PM
Security Audit — agent-trust-hub — pilot-dns