The Agent Skills Directory

[SAFE]: No malicious patterns, unauthorized file access, or credential harvesting were detected. The skill's functionality is limited to reading conversation history and writing to a hardcoded local file (UBIQUITOUS_LANGUAGE.md).
[NO_CODE]: The skill consists purely of natural language instructions and Markdown templates. It does not contain or download any scripts, binaries, or external code.
[PROMPT_INJECTION]: The skill identifies an indirect prompt injection attack surface as it ingests untrusted data from the conversation history.
Ingestion points: Reads the current conversation history (SKILL.md).
Boundary markers: Absent; there are no specific delimiters to separate user conversation from instructions.
Capability inventory: Limited to reading/writing the UBIQUITOUS_LANGUAGE.md file.
Sanitization: Absent; the agent is not instructed to sanitize or escape content from the conversation before writing to the file.
Risk Assessment: The risk is minimal because the skill's capabilities are extremely narrow and the YAML frontmatter explicitly includes disable-model-invocation: true, preventing the agent from calling other tools or services while the skill is active.

ubiquitous-language