The Agent Skills Directory

[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection (Category 8) because it is designed to ingest and process untrusted data from a user's repository, including source code, configuration files, and documentation.
Ingestion points: The skill reads files from the target repository in Phase 2 ("Identify the provider and inventory the workload") and Phase 3 ("Measure the baseline and opportunity").
Boundary markers: No specific boundary markers or instructions to ignore embedded commands are used when reading these external files.
Capability inventory: The agent can execute shell commands via the notte CLI and rg, and it has file-writing capabilities for creating migration branches and implementation areas.
Sanitization: No explicit sanitization or filtering of the read content is mentioned before it is processed by the agent.
[COMMAND_EXECUTION]: The skill utilizes the notte CLI for authentication (notte auth status, notte auth login) and environmental checks. It also uses rg (ripgrep) for searching the local codebase for provider-specific strings such as STEEL_API_KEY or SKYVERN_. These are necessary tools for the skill's stated purpose, and the instructions include a warning to not print or commit any discovered keys.
[EXTERNAL_DOWNLOADS]: The skill references multiple external documentation and pricing URLs for various browser automation providers and benchmarking sites (e.g., Browser Arena, Browserbase, Stagehand). These are documented as legitimate informational references from well-known services and the vendor's own infrastructure.

migrate-to-notte