The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The documentation in SKILL.md contains references to https://book.hacktricks.xyz/pentesting-web, which has been flagged as malicious by automated security scanners. While this site is a common reference for security professionals, the automated detection suggests it may host content deemed risky.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, which introduces a surface for indirect prompt injection.
Ingestion points: The scripts/api_security_tester.py script loads and parses OpenAPI specification files from the local filesystem. Both scripts/api_security_tester.py and scripts/owasp_scanner.py ingest and process HTTP response headers and body content from target URLs provided by the user.
Boundary markers: No explicit boundary markers or instructions are provided to the agent to distinguish external data from its own instructional logic.
Capability inventory: The skill can perform network operations via the requests library and read local files specified as input.
Sanitization: Data is parsed using standard JSON and YAML libraries, but no additional sanitization is applied to protect against malicious payloads within the external content that could influence agent behavior.
[COMMAND_EXECUTION]: The skill provides methodologies and scripts designed to execute security tests by sending payloads (e.g., SQL injection strings, OS command separators) to remote targets. These capabilities are consistent with the skill's primary function but represent inherent risks if misused.

Web Application Security Testing