The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the opencli command-line utility via the Bash tool to perform browser automation, system diagnostics (opencli doctor), and script verification.
[DATA_EXFILTRATION]: The skill provides the agent with capabilities to access and extract sensitive information from the browser context. This includes monitoring and detailing network traffic (opencli browser network), which can expose authentication headers and API responses, and accessing active login sessions via the Strategy.COOKIE mechanism.
[REMOTE_CODE_EXECUTION]: The skill implements a 'sedimentation' workflow where the agent is instructed to generate TypeScript code and write it to the local filesystem (~/.opencli/clis/). This code is then executed using the opencli browser verify command. This dynamic script generation and execution pattern represents a significant capability for running arbitrary code created at runtime.
[PROMPT_INJECTION]: As a tool designed to ingest and process arbitrary web content, the skill is susceptible to indirect prompt injection.
Ingestion points: Web content retrieved via state, get text, eval, and network commands (SKILL.md).
Boundary markers: Absent; there are no specific instructions or delimiters provided to help the agent distinguish between its own instructions and content retrieved from websites.
Capability inventory: The skill has access to shell command execution (Bash), filesystem modification (Write, Edit), and network access (via the browser and fetch in adapters).
Sanitization: Absent; the skill does not define methods for sanitizing or escaping content retrieved from the web before the agent processes it or incorporates it into generated scripts.

opencli-browser