The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches the official Foundry installation script from foundry.paradigm.xyz, which is the industry-standard distribution point for these tools.
[EXTERNAL_DOWNLOADS]: It clones the author's public repository (github.com/pashov/skills) to install the x-ray dependency if it is missing from the environment.
[REMOTE_CODE_EXECUTION]: The environment setup script (ensure_foundry.sh) utilizes a curl | bash pattern to install the Foundry toolchain from its official source.
[COMMAND_EXECUTION]: The skill makes extensive use of local shell commands to build smart contracts, run fuzzers, and execute internal JavaScript helpers for data processing and cost estimation.
[PROMPT_INJECTION]: The workflow involves ingesting untrusted project data, including source code and documentation, which is processed by various LLM subagents with capabilities to write files and execute commands.
Ingestion points: PROPERTIES.md, user-provided documentation, and Solidity source files.
Boundary markers: Not explicitly defined in the subagent prompt templates.
Capability inventory: Includes file writing to test/fizz/, network operations via curl and git, and local command execution via forge, medusa, and echidna.
Sanitization: Relies on standard LLM processing without an additional validation layer for project-supplied content.

fizz