The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted writing samples provided by the user to extract stylistic patterns. While the primary use case is benign, this represents an attack surface for indirect prompt injection if the samples contain malicious instructions aimed at the agent.
Ingestion points: Writing samples of 500+ words each collected in Phase 1.
Boundary markers: The skill does not provide explicit delimiters or instructions to the agent to ignore any embedded directives within the samples.
Capability inventory: The agent has the capability to analyze text and generate new markdown-based skill files.
Sanitization: There is no programmatic sanitization of the input text; the skill relies on manual user filtering.
[DYNAMIC_EXECUTION]: The skill is designed to automatically generate a new SKILL.md file (the voice skill) based on the synthesized analysis of user input. This results in the creation of a new set of agent instructions derived from untrusted content.

voice-analyzer