bbot-recon
Pass
Audited by Gen Agent Trust Hub on Nov 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external network scans, which inherently includes untrusted content from the target or third-party intelligence sources. The instructions explicitly identify this risk and provide a mitigation strategy by directing the agent to treat all fields as data, not instructions.
- Ingestion points: The skill reads JSON events from
bbot-output/<scan_name>/output.json. - Boundary markers: The skill mandates routing events based on
scope_distanceand using specific tools for verification before reporting. - Capability inventory: The skill executes shell commands for the
bbotscanner and reads/parses local JSON output files. - Sanitization: Instructions require the agent to quote external text and treat it as a potential payload rather than actionable instruction.
- [COMMAND_EXECUTION]: The skill constructs and executes shell commands for the
bbotutility. It implements security-positive controls, such as the mandatory--no-depsflag to prevent unauthorized runtime package installations via pip and instructions to pass sensitive API keys through command-line configurations without echoing them to logs.
Audit Metadata