The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a bundled script, scripts/check-availability.sh, using Bash to verify name availability. This script performs read-only checks by querying public APIs and WHOIS records. The skill also utilizes the allowed-tools configuration to limit bash execution to specific commands like whois, curl, and npm view.
[EXTERNAL_DOWNLOADS]: The availability checking process involves querying several well-known and trusted services, including GitHub, npm, PyPI, crates.io, RubyGems, and the WordPress plugin directory. These network operations are strictly for validating the existence of names and do not involve downloading or executing external code.
[PROMPT_INJECTION]: The skill instructions provide a surface for indirect prompt injection (Category 8) by directing the agent to read and analyze project-specific documentation (e.g., in docs/prd/*) to inform the naming strategy.
Ingestion points: User input from the 'Naming Brief' and project files in docs/prd/ or the repository root.
Boundary markers: None specified; the agent is not explicitly warned to treat documentation content as data rather than instructions.
Capability inventory: Network access via curl/whois, local file reading, and package metadata retrieval.
Sanitization: None specified in the instructions. However, this ingestion is the primary intended function of the skill to provide context-aware naming, and no evidence of malicious intent or behavior was found.

naming