The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to use npx graph-spec-cli, which facilitates the download of an external package from the npm registry if it is not already present in the environment.
[REMOTE_CODE_EXECUTION]: By using npx to fetch and run the graph-spec-cli tool, the skill initiates the execution of code from a remote source at runtime.
[COMMAND_EXECUTION]: The skill automates the execution of multiple shell commands such as validate, index, and query. Specifically, the index command includes a --log flag that accepts user-provided strings, creating a potential surface for shell command injection if the input is not strictly sanitized by the agent.
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection as it processes and cross-links markdown concepts from the local filesystem to generate index files and logs.
Ingestion points: Markdown files (<name>.<type-token>.md) and frontmatter within the bundle directory.
Boundary markers: The skill relies on YAML frontmatter structure but lacks explicit instructions to ignore embedded commands or directions within the concept bodies.
Capability inventory: The skill has the ability to write to the filesystem (via the index command) and execute shell commands.
Sanitization: There is no mention of sanitizing or escaping the content of concepts before processing them into the graph or index.

create-graph-spec