The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a Python script to initiate active network connections and TLS handshakes with target servers provided by the user. Active scanning is a sensitive capability that can be misused for unauthorized reconnaissance. The skill implements defense-in-depth via a mandatory authorization gate involving a verbal user prompt and a required CLI flag.
[INDIRECT_PROMPT_INJECTION]: The skill parses and reports metadata from external TLS certificates (e.g., Common Names and Subject Alternative Names) and cipher suite strings. This ingestion of untrusted data from potentially malicious servers creates a surface for indirect prompt injection attacks if the agent processes or displays these strings without further sanitization.
Ingestion points: The _grab_cert_and_cipher function in scripts/analyze_tls.py retrieves raw certificate fields from remote hosts.
Boundary markers: The script generates structured output, but no specific delimiters or warnings are used to signal to the agent that the certificate data should be treated as untrusted content.
Capability inventory: The skill allows file reading and restricted shell execution (Python/OpenSSL).
Sanitization: There is no evidence of sanitization or filtering applied to the certificate metadata to prevent it from containing malicious instructions targeted at the agent.
[DYNAMIC_EXECUTION]: The analyze_tls.py script dynamically modifies sys.path to include the plugin's shared library directory based on its own file location. While standard for this environment, loading modules from computed paths is a dynamic execution pattern that warrants notice.

analyzing-tls-config