context-switcher
ReviewAudited by ClawScan on May 10, 2026.
Overview
The skill is mostly local and purpose-aligned, but its packaged file layout appears inconsistent with its scripts and may cause state files to be written outside the advertised skill folder.
Treat this as needing review before installation. The local-only design is reassuring, but the supplied file layout does not match the documented scripts/modes structure, and the shell scripts may write outside the skill’s own folder. Ask the author to fix or clarify the layout, then review the auto-trigger and DND logging behavior to make sure it matches your privacy and notification preferences.
Findings (4)
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.
The skill may not run as documented, and helpers may read or write files in unexpected locations.
The skill advertises scripts and modes subdirectories, but the supplied manifest/file contents place switch.sh, restore.sh, summarize.sh, and the mode markdown files at the package root. This makes the runnable layout unclear and affects how the scripts resolve paths.
files:
- "scripts/*"
- "modes/*"Fix the package layout or script path logic so the manifest, README, SKILL.md metadata, and runtime paths all agree before installation.
Mode switching could modify shared OpenClaw skill-directory state instead of only this skill’s own files, potentially interfering with other skills or OpenClaw behavior.
Because the provided artifact path is root-level `switch.sh`, this `../` calculation would resolve to the parent of the skill directory, causing `current-context.json` and `snapshots/` to be created or overwritten outside the advertised `context-switcher` folder.
SKILL_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
CONTEXT_FILE="$SKILL_DIR/current-context.json"
SNAPSHOTS_DIR="$SKILL_DIR/snapshots"Only install after the author confirms the intended directory structure, or update the scripts to resolve the actual skill directory without escaping to the parent.
A matching calendar title or phrase could unexpectedly mute notifications or change how OpenClaw responds.
Calendar titles and trigger phrases can automatically change the assistant’s mode, response style, and notification behavior. This is disclosed and aligned with the skill’s purpose, but it affects agent behavior without a separate approval each time.
This skill may be invoked autonomously by OpenClaw when a trigger phrase is detected in your messages, or when a calendar event title matches a known mode keyword.
Review the trigger terms and disable auto-triggering if you do not want calendar events or casual phrasing to switch modes automatically.
Sensitive messages received during DND may remain in a local file until restored or cleared.
DND mode persists incoming message/task content locally for later summary. This is disclosed and purpose-aligned, but the log may contain sensitive private or work information.
Log all incoming messages and tasks silently to `snapshots/dnd-log.json` for review on exit
Use DND logging only if local storage of missed messages is acceptable, and periodically inspect or delete the snapshot files.