The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run search commands like rg on project files and sensitive local configuration paths, specifically ~/.claude/settings.json and ~/.claude/skills/, to find existing capabilities.
[EXTERNAL_DOWNLOADS]: The workflow describes identifying and installing third-party packages from registries like npm and PyPI. It references numerous well-known tools (e.g., prettier, zod, pydantic) as established solutions for adoption.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes data from untrusted external repositories and websites. \n
Ingestion points: External search results from npm, PyPI, GitHub, and general web sources. \n
Boundary markers: Absent. The instructions do not define delimiters or provide warnings to ignore embedded instructions in external content. \n
Capability inventory: The agent can execute shell commands (rg, package managers) and perform file system operations. \n
Sanitization: Absent. No methods for validating or escaping external content before interpolation are described.

search-first