The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill fetches documentation from external sources via MCP tools to guide the agent's code generation, which introduces a surface for potentially untrusted data to influence agent behavior.
Ingestion points: Documentation content fetched via search_arcium_docs and query_docs_filesystem_arcium_docs as specified in SKILL.md.
Boundary markers: The skill does not define explicit delimiters or instructions for the agent to ignore commands that might be embedded within the retrieved documentation.
Capability inventory: The skill facilitates the creation and building of Solana programs and Arcis circuits using the arcium CLI and Anchor framework.
Sanitization: There is no documented mechanism for validating or sanitizing the external documentation content before it is integrated into the agent's context.
[REMOTE_CODE_EXECUTION]: The skill provides instructions for installing the Arcium CLI using a remote script.
Evidence: curl -sSfL https://install.arcium.com/ | bash in references/troubleshooting.md.
Context: The script originates from the vendor's official domain (arcium.com) and represents a standard distribution method for their development tools.
[COMMAND_EXECUTION]: The skill requires the agent to perform several command-line operations to manage the application development lifecycle.
Tools: arcium (commands: build, test, deploy, clean, localnet), anchor (commands: build), and docker for environment orchestration.
Scope: These commands are used appropriately within the scope of building and testing encrypted computations on the Solana blockchain.

arcium