The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx -y to fetch the utilia-solana-agent package from the npm registry. This occurs at runtime and introduces a dependency on an external package not managed within the skill's own files.
[REMOTE_CODE_EXECUTION]: By using npx -y utilia-solana-agent, the skill executes code downloaded from a remote repository. The use of -y skips confirmation prompts, and the absence of a specific version tag means the agent always executes the latest version of the utility, which could be subject to supply-chain attacks.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands that interact with the local filesystem (reading keypairs) and the Solana network. This involves managing environment variables like SOLANA_PRIVATE_KEY which contains highly sensitive cryptographic material.
[PROMPT_INJECTION]: The skill processes untrusted content from external PDF URLs and converts it into Markdown for the agent to use in downstream tasks, creating a surface for indirect prompt injection.
Ingestion points: Public PDF URLs processed via the pdf-to-markdown command in SKILL.md.
Boundary markers: Absent; there are no delimiters or instructions provided to the agent to treat the converted Markdown as untrusted data.
Capability inventory: The skill environment has access to shell execution, network operations, and sensitive environment variables (Solana private keys).
Sanitization: Absent; the skill returns the raw extracted text to the agent for processing.

utilia-pdf-to-markdown