automation-recon-pipeline

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides shell command examples for chaining security tools (subfinder, dnsx, httpx, nuclei) into a pipeline. These are standard security research practices.
  • [PERSISTENCE]: The documentation suggests using standard system scheduling tools like cron and systemd-timer for continuous monitoring. In the context of the skill's stated purpose (recon automation), this is a legitimate workflow suggestion rather than a malicious persistence attempt.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes a workflow that ingests untrusted external data (discovered subdomains and host response headers).
  • Ingestion points: Tools like subfinder and httpx ingest data from external network sources into local files (e.g., live.txt).
  • Boundary markers: None explicitly defined in the example shell pipes.
  • Capability inventory: The pipeline executes several binary tools (nuclei, httpx, etc.) which perform network operations and file writes.
  • Sanitization: Relies on the individual security tools' internal handling of malformed or malicious network responses.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 11:01 PM
Security Audit — agent-trust-hub — automation-recon-pipeline