The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands including rg (ripgrep), grep, awk, find, and comm to analyze local source code and identify function definitions and access modifiers.
[EXTERNAL_DOWNLOADS]: The documentation suggests installing well-known security tools such as slither-analyzer and solc-select via pip, as well as system packages like ripgrep via apt or brew.
[INDIRECT_PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests local source code for analysis.
Ingestion points: Files within the directory specified by SECSKILL_TARGET (Solidity, Rust, Move, or web API source code).
Boundary markers: None identified in the workflow.
Capability inventory: Uses subprocesses to run rg, grep, slither, and solc-select for static analysis.
Sanitization: The skill uses regex-based filtering (rg, grep) and specialized static analysis tools rather than feeding arbitrary file content directly into the agent's reasoning loop for instruction execution.

entry-point-analyzer