The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill uses uvx to fetch and execute code directly from the external repository https://github.com/ambient-code/agentready. This involves downloading and running unverified scripts from an external source at runtime.
[COMMAND_EXECUTION]: The skill extracts and executes shell commands from a JSON report (assessment-latest.json) produced by the external tool. By following the remediation.commands field without validation, the skill allows an external process to define the commands executed on the user's system.
[DATA_EXFILTRATION]: The skill reads the file ~/.claude/skills/rhdh/references/rhdh-repos.md. This is a vendor-specific configuration file used to provide context for repositories within the Red Hat Developer Hub (RHDH) ecosystem.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection via the scan results it processes.
Ingestion points: The skill parses the report assessment-latest.json, which contains data generated from scanning the contents of the local repository.
Boundary markers: No delimiters or safety instructions are used to distinguish the tool's output from trusted logic.
Capability inventory: The skill can execute shell commands and modify files based on the remediation metadata in the report.
Sanitization: No validation or sanitization is performed on the commands or remediation steps extracted from the JSON file before execution.

agent-ready