The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional text for rephrasing conversational content. It does not contain scripts, binaries, or shell commands.
[SAFE]: References to external resources (such as the ASD-STE100 standard website) are provided solely for documentation and background context, with no automated fetching or execution of remote content.
[INDIRECT_PROMPT_INJECTION]: The skill monitors user input for signals of confusion to trigger its re-explanation logic. While this is an ingestion surface for untrusted data, the risk is negligible as the skill only generates conversational text and lacks capabilities to write files, execute code, or perform network requests.
Ingestion points: User signals of non-comprehension (e.g., 'wait, what') and previous conversational context (SKILL.md).
Boundary markers: Not explicitly defined, relying on standard agent message handling.
Capability inventory: Reading local repository documentation files like README.md and CONTEXT.md (SKILL.md).
Sanitization: No specific sanitization of user signals is performed as they are only used as triggers for rephrasing.

re-pitch