The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions to execute various development tools including pnpm, vitest, openclaw CLI, and curl for local gateway testing. These commands are necessary for the skill's primary function of end-to-end verification.
[DATA_EXFILTRATION]: Instructions include reading local configuration files at ~/.openclaw/openclaw.json and session metadata. This involves accessing application-specific data to facilitate testing and debugging of the gateway stack.
[PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it ingests and processes session transcripts that may contain untrusted user input from connected channels.
Ingestion points: Reads session transcript files (.jsonl) and session metadata (sessions.json) from the ~/.openclaw directory.
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potentially malicious content within these logs.
Capability inventory: The skill allows for shell command execution (pnpm, openclaw, python3), local network requests (curl), and file system operations.
Sanitization: There is no evidence of sanitization or filtering applied to the transcript content before the agent processes it for verification purposes.

nexu-e2e-test