The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a shell command to perform project scaffolding.
Evidence: Calls npx create-agent-harness <name> in the 'Under the hood' section of SKILL.md.
[EXTERNAL_DOWNLOADS]: The skill utilizes npx to download and run the create-agent-harness package from the npm registry.
Evidence: Use of npx create-agent-harness in SKILL.md.
[PROMPT_INJECTION]: User-provided data, such as the harness name, is interpolated into a shell command, creating a surface for potential indirect prompt injection or command injection.
Ingestion points: User input for 'Harness name' and 'Description' in SKILL.md.
Boundary markers: Absent; instructions do not specify validation or delimiters for the name variable to prevent shell metacharacter injection.
Capability inventory: Shell command execution via npx in SKILL.md.
Sanitization: Absent; the skill relies on the agent's default handling of variables in shell commands.

create-harness