The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes shell scripts (scripts/audit-target.sh and scripts/validate.sh) that the agent is instructed to execute. audit-target.sh uses system tools like find, rg (ripgrep), and cargo to analyze target projects. validate.sh executes a local Python validation script. These are utility tools for the agent's primary purpose and do not appear to perform malicious actions.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to audit and refactor external Rust projects, which creates a surface for indirect prompt injection if the ingested code contains malicious instructions.
Ingestion points: scripts/audit-target.sh ingests file paths, dependency trees, and source code content from a user-provided directory.
Boundary markers: No explicit boundary markers or instructions to ignore embedded prompts are provided in the script or instructions.
Capability inventory: The skill uses shell commands (rg, find, cargo tree) to discover project structures and dependencies.
Sanitization: No sanitization or filtering of the analyzed source code content is performed by the discovery script.

asupersync-mega-skill