The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read a file named CONTEXT.md from the user's codebase to ensure domain-specific naming conventions are followed. This represents an ingestion point for untrusted data that could potentially contain malicious instructions.
Ingestion points: CONTEXT.md (referenced in SKILL.md).
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat the content of CONTEXT.md as data rather than instructions.
Capability inventory: The skill is entirely documentation-based and configuration in SKILL.md explicitly includes disable-model-invocation: true, which restricts the agent's ability to trigger tools while using this skill.
Sanitization: There are no mechanisms described for validating or sanitizing the content of the referenced file.
[NO_CODE]: The skill consists solely of Markdown documentation and YAML configuration files. No scripts (.py, .js, .sh), binaries, or executable files are distributed with the skill.

tdd