The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (MEDIUM): The skill guides users to install the entire CLI via Homebrew or Go from an external repository (github.com/entireio/cli). This source is not on the pre-approved list of trusted organizations, requiring manual verification of the source's integrity.
[PROMPT_INJECTION] (HIGH): The skill exhibits a significant Indirect Prompt Injection surface (Category 8).
Ingestion points: The agent processes session transcripts, AI prompts, and historical responses via the entire explain command and reads external repository states during entire rewind operations.
Boundary markers: There are no specified delimiters or instructions to ignore embedded commands within the transcripts being analyzed.
Capability inventory: The skill possesses high-privilege capabilities including direct filesystem modification (code restoration), git branch management, and destructive operations like entire reset --force.
Sanitization: There is no evidence of sanitization or filtering of the transcript content before it is re-introduced into the agent's reasoning context.
[COMMAND_EXECUTION] (LOW): The skill utilizes a suite of CLI commands (entire enable, entire status, entire rewind, entire doctor) to manage project state. While these are intended functions, they grant the agent the ability to execute complex shell operations that could be abused if the agent is subverted.

entire