Self Reflection
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only skill is purpose-aligned and disclosed, but it does create and reuse persistent reflection memory that can influence future agent behavior.
This skill appears safe and coherent for self-review. Before installing, be comfortable with the agent keeping local reflection notes under ~/reflection/ and optionally registering itself in workspace memory or heartbeat files. Periodically review those files and remove sensitive or stale lessons.
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.
Old, incorrect, overly broad, or sensitive notes could bias future responses or reveal prior project context in later sessions.
The skill intentionally stores and reuses prior reflections to influence future agent behavior.
Before significant work, scan `~/reflection/patterns.md`. Surface relevant lessons
Review and edit ~/reflection/patterns.md and ~/reflection/reflections.md periodically, and avoid logging sensitive details unless you want them reused later.
The agent may continue applying this reflection process across sessions until the memory entry is removed.
The setup instructions make the skill persist as an active workspace memory entry.
Register in workspace memory — Add to MEMORY.md ... Status: active
Install the workspace memory entry only if you want ongoing behavior changes, and remove it from MEMORY.md if you no longer want the skill active.
If enabled, the agent may periodically check and update reflection state as part of normal heartbeat behavior.
The skill documents an optional recurring review trigger, which is disclosed and aligned with its purpose.
Heartbeat (Optional) ... Every heartbeat, quick scan: 1. Pending reflection to log? 2. Pattern threshold reached
Enable the heartbeat integration only if you want recurring reflection checks; otherwise leave HEARTBEAT.md unchanged.