The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines a PostToolUse hook that executes a shell script at ../../../hooks/skills/post_output_validation_hook.sh after the Read tool is invoked. This script is located outside the skill's immediate directory, though it likely belongs to the vendor's shared environment.
[PROMPT_INJECTION]: The skill processes untrusted data from multiple external sources, creating a surface for indirect prompt injection.
Ingestion points: Data is ingested from javascript_signals, html_signals, dns_signals, http_signals, and job_signals.
Boundary markers: There are no explicit delimiters or instructions provided to isolate the untrusted signal data from the agent's operational instructions.
Capability inventory: The skill utilizes Read and Grep tools and executes an external command-line validation hook.
Sanitization: No sanitization, filtering, or validation logic is defined for the external signals processed by the skill.

third-party-detector