pilot-recruitment-pipeline-setup
Pass
Audited by Gen Agent Trust Hub on Jun 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the Bash tool to run various CLI commands, including host configuration and agent synchronization. It also creates a local configuration directory at
~/.pilot/setups/and writes JSON manifest files using shell commands. - [EXTERNAL_DOWNLOADS]: The skill uses the
clawhubutility to download and install several functional components (e.g.,pilot-discover,pilot-metrics,pilot-alert) from the Pilot Protocol ecosystem. - [PROMPT_INJECTION]: The skill creates a pipeline that ingests untrusted data from external sources, specifically candidate profiles from job boards and LinkedIn, which creates a risk of indirect prompt injection.
- Ingestion points: Candidate profiles are retrieved from external websites and social networks via the 'sourcer' agent.
- Boundary markers: The manifests do not specify delimiters to separate untrusted candidate content from system instructions.
- Capability inventory: The system has the capability to publish data between agents and send data to external webhooks (port 443).
- Sanitization: No explicit sanitization or validation of external candidate data is mentioned in the setup procedure or manifests.
Audit Metadata