Self Improvement
PassAudited by ClawScan on May 10, 2026.
Overview
This is a disclosed learning-log and reminder skill, but users should curate what it stores or shares because entries can persist across future agent sessions.
Install this only if you want persistent learning memory for your agent. Before enabling hooks, review the scripts; avoid logging secrets or full sensitive command outputs; and manually approve anything promoted to long-term instruction files or shared with other sessions.
Findings (3)
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.
Future agent sessions may be influenced by anything logged or promoted, including stale corrections, command-output details, or accidental secrets.
The skill is designed to reuse learning files and promoted guidance in future sessions; this is its stated purpose, but it means incorrect, prompt-like, or sensitive entries can persist.
OpenClaw injects these files into every session: ... AGENTS.md ... SOUL.md ... TOOLS.md ... MEMORY.md ... .learnings/
Review and sanitize .learnings entries, avoid storing secrets or full sensitive outputs, and require user review before promoting entries into long-term prompt or memory files.
If used carelessly, learnings or transcript details from one session could be exposed to another session or background agent.
The skill documents cross-session transcript access, messaging, and sub-agent spawning as ways to share learnings. This is disclosed, but session identity and data-boundary details are not described in the skill.
sessions_history — Read another session's transcript ... sessions_send — Send a learning to another session ... sessions_spawn — Spawn a sub-agent for background work
Only share sanitized, necessary snippets; confirm the target session before sending; and avoid reading other transcripts or spawning sub-agents unless the user intends it.
Enabling hooks means local code from the skill can run automatically during agent activity, even though the included scripts are simple reminders.
Optional hook setup can execute local helper scripts during prompt or tool-use events. The provided scripts appear scoped to reminder output, and the documentation discloses that enabling them is opt-in.
Hook scripts run with the same permissions as Claude Code ... Scripts only output text; they don't modify files or run commands ... All scripts are opt-in
Inspect the hook scripts before enabling them, keep hook setup opt-in, and disable the hook if you no longer want persistent reminders.