The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches installation scripts and configuration from the official Polylane domain (polylane.com) and the @coreplane/polylane package.
[REMOTE_CODE_EXECUTION]: Provides commands to download and execute scripts directly in the shell (curl | bash and irm | iex) from the vendor's web domain to install the CLI utility.
[COMMAND_EXECUTION]: Utilizes a dedicated CLI tool to perform system-level operations, including cloud resource exploration, log retrieval, and service management.
[DYNAMIC_EXECUTION]: Supports the execution of arbitrary JavaScript code snippets via the polylane tools code command, allowing users to script interactions with the platform's internal toolset.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes content from external sources that may be controlled by third parties, creating a potential path for indirect instruction injection.
Ingestion points: The agent reads logs (polylane service logs), repository search results (polylane repo grep), and issue trackers (polylane issue show) in SKILL.md.
Boundary markers: There are no specific instructions or delimiters provided to prevent the agent from following instructions embedded within the ingested data.
Capability inventory: The skill has high-privilege capabilities, including the ability to connect cloud accounts (polylane cloud connect), execute platform tools (polylane tools run), and run custom code.
Sanitization: No evidence of validation or sanitization of the retrieved external data is present in the skill instructions.

polylane-cli