The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides bash command templates for the agent to search and analyze local conversation logs. These commands utilize standard utilities like jq, ripgrep (rg), and awk to filter and format JSONL data within the agent's storage directory (~/.openclaw/agents/).\n- [SAFE]: No network operations, external downloads, or remote code execution patterns were identified. All data processing occurs locally on the agent's history files.\n- [SAFE]: No hardcoded secrets, obfuscation, or privilege escalation attempts were detected. The skill accesses application-specific data rather than sensitive system configuration files.\n- [SAFE]: (Indirect Prompt Injection) The skill ingests untrusted historical data from session logs. While no explicit boundary markers or sanitization logic is provided in the bash snippets, the operations are limited to text processing and do not interact with dangerous system capabilities.

session-logs