Context Pruner
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only skill is coherent and low risk, but it can summarize and discard conversation context, so important details should be reviewed before pruning.
This skill appears safe to install as an instruction-only context-management helper. Be aware that it may condense or remove parts of a conversation, so review generated summaries before using them for tasks where exact wording or full history matters.
Findings (1)
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 wording, caveats, or user preferences could be lost if the agent prunes context too aggressively.
The skill intentionally changes the context the agent retains by summarizing and discarding parts of a conversation. That is aligned with its purpose, but summaries can omit nuance or become over-trusted in later turns.
**Fact Distillation**: Extract raw info and discard the conversational fluff. ... **Chunking**: Break long transcripts into searchable summaries.
Use this skill for long sessions, but ask the agent to show or confirm summaries before relying on them for important decisions.