Announcing Beacon: Self-improving memory for coding agents
← All skills
Skill

performing-web-application-scanning-with-nikto

skill:mukul975/anthropic-cybersecurity-skills#performing-web-application-scanning-with-nikto

Runs Nikto, an open-source web server and web application scanner, to test over 7,000 potentially dangerous files/programs, check for outdated versions across 1,250+ servers, and identify XSS, SQL injection, misconfigurations, default credentials, and vulnerable CGI scripts. Use when performing web application vulnerability scanning, security assessments, scheduled security testing, or validating web server security controls.

Verdict
Suspicious

Risky behavior the analysis could not rule out.

Analyzed Oct 27, 2026, 23:35 UTC at 54a7988

Dispute security verdict

If you believe this verdict is incorrect, submit a dispute and we'll review it.

1
findings
high
highest severity
1
threat classes
4,532
files scanned
119/119
checks satisfied
0
package advisories

Why this verdict

  • suspicious: finding skill:skills/performing-web-application-scanning-with-nikto:skills/performing-web-application-scanning-with-nikto/scripts/process.py:39:DESTRUCT

Security findings 1

Destructive action

highContextual risklow confidence

The bundled process.py scanner enables Nikto's Denial of Service test class (tuning id 6, enumerated in SKILL.md) by default via tuning='123456789abc', so every scan performs availability-impacting probes against the designated targets without explicit DoS opt-in.

tuning: str = 
vector:code
Counter-argument considered

The shipped automation script's default tuning='123456789abc' embeds Nikto tuning id 6, which SKILL.md itself enumerates as 'Denial of Service', and this value flows unmodified into the subprocess.run nikto command, so the availability-impacting probe class is on by default; the declared description enumerates only non-availability test types (XSS, SQLi, misconfig, default creds, CGI) and declared claims are empty, so no carve-out (including declared core capability) covers it and the low-confidence contextual_risk finding stands.