The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use the rnd binary for project auditing and package queries. It includes a specific safety instruction that the agent should not install the binary itself, keeping that responsibility with the user.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes data from external sources and local files. 1. Ingestion points: The skill fetches data from the React Native Directory API and scans local source code via rnd analyze. 2. Boundary markers: Instructions specify that the CLI emits structured JSON facts and the agent should reason about the data independently. 3. Capability inventory: The agent has the capability to execute shell commands (specifically the rnd tool). 4. Sanitization: There is no evidence of sanitization or validation performed on the data retrieved from the external API.
[COMMAND_EXECUTION]: The instructions in SKILL.md contain hardcoded absolute paths (/Users/galhavkin/Desktop/...) which expose details of the developer's local file system and may lead the agent to attempt accessing non-existent directories.

rnd-cli