The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes routine Docker and Yarn commands (docker info, yarn workspace ... docker:build, yarn workspace ... docker:compose) to set up and run a local testing environment. These are standard operations for development and testing.
[DATA_EXPOSURE]: The skill instructs the agent to read local test results and service logs (test-results.json, *.log) from the project's log directory. This access is restricted to the local workspace and is used solely for failure diagnosis.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes test outputs and logs at runtime.
Ingestion points: Reads test-results.json and service logs from packages/evm-wallet-experiment/logs/.
Boundary markers: None present.
Capability inventory: Executes docker and yarn commands.
Sanitization: None present.
Context: This is considered safe as the data being processed is generated locally by the project's own test suite during execution.

evm-wallet-docker-e2e