The Agent Skills Directory

[SAFE]: The skill does not perform any network operations, file system modifications (beyond manual setup documentation), or credential access. All instructions are focused on response formatting and user accessibility.
[PROMPT_INJECTION]: The skill includes instructions to maintain a specific formatting style throughout a session, regardless of topic changes. While this defines a persistent behavioral persona, it explicitly defers to the platform's system prompt and safety guidelines in case of conflict, ensuring core security constraints remain active.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a wrapper for user-provided data.
Ingestion points: The {{args}} placeholder in agents/gemini.toml ingests user-supplied content.
Boundary markers: Explicit delimiters for the injected content are absent in the prompt template.
Capability inventory: The skill possesses no dangerous capabilities; it does not invoke subprocesses, perform file writes, or initiate network operations.
Sanitization: No escaping or filtering of user content is performed before interpolation.
[DYNAMIC_CONTEXT_INJECTION]: No use of shell execution syntax (exclamation mark + backticks) was found within the skill's instructional content.

i-have-adhd