Session Health Monitor
PassAudited by ClawScan on May 1, 2026.
Overview
The skill appears purpose-aligned, but it deliberately writes session facts to local memory and can install a persistent local statusline command.
This skill is reasonable for context-health monitoring. Before installing, decide whether you want persistent memory snapshots and a Claude statusline command, review the snapshot contents for sensitive information, and be aware that rotation archives old memory files rather than deleting them.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Important session details may remain in local agent memory and could influence or reappear in later work.
The skill stores selected session facts in a persistent local memory file; this supports the stated snapshot purpose, but retained facts can include sensitive or inaccurate context that later agents may reuse.
daily_file="$memory_dir/${today}.md" ... echo "## Pre-Compaction Snapshot ($time_now)" ... echo "- $fact" >> "$daily_file"Review what gets snapshotted, avoid storing secrets, set MEMORY_DIR deliberately, and periodically review or delete archived memory files.
After setup, Claude Code can run the statusline shell script automatically when rendering the status line.
The setup script installs a local shell script and configures Claude's statusLine to run it; this is disclosed and central to the statusline feature, but it creates ongoing local command execution.
cp "$SCRIPT_DIR/statusline.sh" "$STATUSLINE_DEST" ... jq --argjson sl "$statusline_config" '.statusLine = $sl' "$SETTINGS_FILE"
Run the setup script only if you want the persistent statusline behavior; inspect the script first and keep or use the generated settings backup if you need to revert.