session-recover

What This Skill Does

The session-recover skill is an essential recovery tool for OpenClaw agents, designed to restore context and continuity after an unexpected session reset or when an agent simply needs to review the history of a previous engagement. By interfacing with local JSONL session logs, it allows the agent to parse, summarize, and extract critical information from past dialogues. It does not rely on transient memory hooks; instead, it performs direct forensic analysis on saved session data, ensuring you never lose track of technical decisions, pending tasks, or nuanced conclusions reached in earlier sessions. This makes it an indispensable utility for maintaining long-term progress across fragmented conversations.

Installation

To install this skill, use the ClawHub command within your OpenClaw terminal: clawhub install openclaw/skills/skills/axelhu/session-recover

Use Cases

• Context Recovery: When a session is accidentally reset, use this skill to quickly retrieve the preceding state and continue from the last successful checkpoint.
• Session Summarization: Perfect for long-running projects; trigger a report to consolidate the dialogue into actionable insights, key conclusions, and pending items.
• Historical Auditing: Search through previous sessions using specific keywords to find technical configurations, code snippets, or decisions made weeks ago.
• Progress Tracking: Maintain momentum by extracting pending items and "next steps" from archives, preventing tasks from falling through the cracks.

Example Prompts

• "/recover – can you check the last 20 messages of this session to remind me where we left off with the API integration?"
• "/recover – search the current session logs for 'database migration' and pull out any pending steps we haven't finished yet."
• "/recover – I think our previous session was reset; please look for the most recent .reset file and summarize the key technical points we discussed."

Tips & Limitations

• Log Parsing: The effectiveness of this skill depends on the availability of local JSONL files. Ensure your agent's sessions directory is correctly mapped.
• Efficiency: Use the --tail flag for quick check-ins, and the --keyword flag to filter through large session logs quickly. This prevents unnecessary overhead.
• Data Privacy: Because this skill reads local file system logs, ensure that sensitive info in those logs is stored according to your local security protocols, as the script will process everything contained within the JSONL structures.