Browser Bookmarks Cleanup

Analyze, organize, and clean browser bookmarks on macOS using on-disk bookmark and history files. Detects duplicates, stale bookmarks, tracking parameters, a...

Security Scan

The skill's code, instructions, and requirements are consistent with a local macOS browser-bookmarks analysis and cleanup tool and do not request unrelated credentials or network access.

Purpose & Capability

Name and description match the included scripts: discovery of Chrome/Firefox profiles, read-only analysis of bookmarks/history, proposal of a JSON plan, and an opt-in apply step that creates backups. The declared scope (macOS browser files) is consistent with the hardcoded BROWSER_ROOTS and commands.

Instruction Scope

SKILL.md commands map directly to the CLI scripts. The runtime instructions confine operations to local files (Bookmarks JSON, places.sqlite, History) and require explicit user approval and a --write flag for changes. The SKILL.md 'Hard Safety Rule' (no network requests to bookmark URLs) is respected by the code (no network libraries or outbound calls present).

Install Mechanism

There is no install spec or external download; the skill is delivered as code files that run with the system Python. No third-party packages, remote URLs, or extract/install steps are present in the package.

Credentials

The skill requests no environment variables, credentials, or external tokens. It accesses only local browser files (Bookmarks JSON, Local State, places.sqlite, History) appropriate for its stated purpose.

Persistence & Privilege

The skill is not always-enabled and is user-invocable. It does not modify other skills or system-wide agent settings. Write operations are opt-in, perform a timestamped backup, and the code provides a rollback path.

Assessment

This skill appears to do only local analysis and opt-in cleanup of browser bookmarks and history. Before running: (1) review the generated plan JSON before using --write, (2) run a dry-run and keep the backup files it creates, (3) close the browser before committing changes as recommended, and (4) be aware the tool reads your History (places.sqlite) which contains visited URLs — treat that data as sensitive. Note the skill package has no homepage and an unknown publisher; if you require provenance, ask the publisher for source hosting or a reproducible release before trusting it with your real profiles.

Browser Bookmark Cleanup

Analyze first, organize, then clean up. All writes are opt-in and reversible.

Supported browsers: Chrome and Firefox out of the box. Other Chromium-based browsers (Edge, Brave, Arc, etc.) can be handled by pointing to their profile directories, which use the same Bookmarks JSON format as Chrome.

Workflow

1. Discover browser profiles.
2. Run read-only analysis and share findings.
3. Propose cleanup plan with preview.
4. Get explicit user approval before any writes.
5. Dry-run first, then write with automatic backup.
6. Report changes and how to roll back.

Hard Safety Rule

Never send network requests to bookmark URLs. Analysis uses only local files.

Commands

Discover profiles:

python3 scripts/browser_bookmarks.py discover [--browser chrome|firefox]

Analyze bookmarks:

python3 scripts/browser_bookmarks.py analyze --bookmarks "<PATH>" --output /tmp/bookmark-analysis.json

Apply plan (dry-run by default, --write to commit):

python3 scripts/browser_bookmarks.py apply-plan --bookmarks "<PATH>" --plan /tmp/bookmark-plan.json [--write]

Analysis Categories

1. Exact and semantic duplicate URLs
2. Tracking-parameter variants (utm_*, gclid, fbclid, etc.)
3. Subdomain and domain concentration
4. HTTP links that should be HTTPS
5. Empty, singleton, deep, and oversized folders
6. Weak bookmark names
7. Old bookmarks by age
8. Never-visited and stale bookmarks (from browser history)

Approval Gate

Before any write: show plan preview, get explicit approval, ask user to close the browser.

Plan Format

See references/plan-schema.md for the JSON plan structure.

Rollback

Restore the timestamped backup file created before any write.