The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes git remote -v to identify the project's remote origin host.
[PROMPT_INJECTION]: The instructions contain directives to bypass human-in-the-loop oversight, such as "No confirmation rounds", "take the recommended default for every section — derived from what the repo itself implies — then write", and "summarise... then take each section's recommended default without asking" (SKILL.md). This encourages the agent to perform repository modifications without user verification.
[INDIRECT_PROMPT_INJECTION]: The skill generates persistent configuration that influences the behavior of other agent skills.
Ingestion points: Reads .git/config, AGENTS.md, CLAUDE.md, CONTEXT.md, CONTEXT-MAP.md, and package.json (SKILL.md).
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore or sanitize embedded directives within the files it reads.
Capability inventory: Performs command execution (git), file system reads, and multiple file system writes (modifying CLAUDE.md/AGENTS.md and creating files in docs/agents/) (SKILL.md).
Sanitization: The skill does not specify any sanitization or validation process for the data extracted from repository files before it is used to generate new instructions.

setup-utarn-skills-yes