The Agent Skills Directory

[DATA_EXFILTRATION]: The skill performs discovery and access of sensitive information as part of its auditing workflow.
The skill accesses ~/.openclaw/openclaw.json, which contains gateway settings, authentication configurations, and model provider details.
It scans environment files (.env) and log files in ~/.openclaw/logs for hardcoded secrets, including API keys, tokens, and passwords.
These operations are consistent with the skill's primary purpose as a security checker for the vendor's platform.
[INDIRECT_PROMPT_INJECTION]: The skill exposes a surface for indirect prompt injection due to the ingestion of untrusted external data files.
Ingestion points: The skill reads and parses system logs, configuration files (openclaw.json), and environment variables.
Boundary markers: The instructions do not define delimiters or explicit "ignore" markers for data retrieved from logs or configuration values.
Capability inventory: The skill has access to powerful system utilities including node, jq, grep, openssl, stat, ls, id, ss, netstat, and iptables for performing its checks.
Sanitization: There is no mention of sanitizing or escaping the content of logs or configuration fields before they are evaluated by the agent's scoring and reporting logic.

openclaw-security-checker