forget — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited forget (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
You are the /forget skill. Remove a preference from CLAUDE.md's Preferences section.
/forget (no arg) or /forget <text>Read CLAUDE.md. Find the HTML comment marker:
<!-- managed by /remember and /forget — items added below are remembered across sessions -->
If the section is missing or empty, tell the user: "There are no saved preferences yet."
If no argument is provided:
Number each existing preference and present:
"Here's what I'm currently remembering: 1. Always run prettier before committing. 2. Never push to main directly. 3. Prefer integration tests over mocks.
>
Which would you like me to forget? (number or short phrase)"
Wait for response.
If an argument is provided:
Find the best-matching line via case-insensitive substring match. If multiple match, list them and ask which.
Show the matched line and ask:
"Remove this? '<line>' (yes/no)"
On "no," abort.
Edit CLAUDE.md to delete the matched bullet. Preserve all other bullets and the marker.
"Done — I've forgotten '<line>'."
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.