The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides complex shell instruction blocks in references/validation.md intended for direct execution by the agent.
Evidence: Instructions use npm test, npm run build, and the pi CLI tool to perform installations and validation loops.
Evidence: The validation workflow involves file system manipulation, including mktemp -d for environment setup and rm -rf for cleanup.
[COMMAND_EXECUTION]: Dynamic script execution via heredocs (hereditary documents).
Evidence: In references/validation.md, inline Python code is piped to python3 - to extract package names and validate manifests from package.json files.
[PROMPT_INJECTION]: Indirect prompt injection vulnerability surface.
Ingestion points: The skill is designed to read and process external package.json files and other project resources (references/validation.md).
Boundary markers: Absent. There are no specified delimiters or instructions for the agent to ignore embedded commands within the data being processed.
Capability inventory: The skill utilizes high-privilege capabilities including shell command execution, file system writes, and package manager operations (references/validation.md).
Sanitization: Absent. The logic does not specify sanitization or escaping for data interpolated from the package manifests into the shell commands or agent context.

pi-package-creator