The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local binary located at ./target/debug/wt to perform configuration, workflow creation, and execution tasks.
[COMMAND_EXECUTION]: Uses common shell utilities including git, find, and sed to inspect the repository state and read project documentation.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by processing external data from the local repository.
Ingestion points: Reads data from docs/consistency.md, wt/ideas/, wt/tasks/, and wt/workflows/ to determine experiment shapes and task goals.
Boundary markers: No explicit boundary markers or 'ignore' instructions are provided when reading these files.
Capability inventory: The skill can write files to the filesystem (wt/tasks/, wt/profiles/) and execute the wt binary with parameters derived from its analysis.
Sanitization: There is no evidence of sanitization for the content ingested from the project files before it is used to construct new prompts or commands.

wt-variants