Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Delete a memory from Origin by ID. Destructive and cannot be undone — prefer `/capture` with `supersedes` for corrections. Invoked as `/forget <source_id>`.
.claude/skills/davepoon-forget/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 451% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -54% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -50% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -67% | 0% |
Permanently delete a memory by its source_id.
You need the source_id. If the user did not provide it, call /recall first to find the matching memory and confirm with the user before deleting.
forget(memory_id="<source_id>")supersedes pointingat the old one. That preserves history. Use /capture with the supersedes arg instead.
/review first, confirm with the user,then delete one at a time.
Deletion is destructive. Always confirm with the user before calling forget, unless the user has already given an explicit, unambiguous instruction in the same turn (e.g. they pasted the ID and said "delete this").
If the deletion removed a page md (daemon archives the page and KnowledgeWriter unlinks the file), snapshot the change. Defensive — silent skip if git missing, ~/.origin/ not a repo, or no diff.
Bash: git -C ~/.origin add -A && \
git -C ~/.origin -c user.name=Origin -c user.email=daemon@origin.local \
commit --quiet -m "forget: <source_id>" 2>/dev/null || \
(sleep 1 && git -C ~/.origin add -A && \
git -C ~/.origin -c user.name=Origin -c user.email=daemon@origin.local \
commit --quiet -m "forget: <source_id>" 2>/dev/null) || trueThe retry handles index.lock races — the daemon may be writing to ~/.origin/ at the same moment (auto-commit from captures). One-second wait is enough for the daemon to release the lock.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→fail | 4,634 | 8,415 | +82% | 1 | 1 | 0% | 567 | 879 | +55% | 0 | 0 | — |
case-01 | fail→fail | 6,089 | 9,185 | +51% | 1 | 1 | 0% | 836 | 860 | +3% | 0 | 0 | — |
case-02 | fail→fail | 5,707 | 10,025 | +76% | 1 | 1 | 0% | 878 | 988 | +13% | 0 | 0 | — |
case-04 | fail→fail | 2,589 | 19,679 | +660% | 1 | 1 | 0% | 245 | 2,412 | +884% | 0 | 0 | — |
case-05 | fail→fail | 1,753 | 13,632 | +678% | 1 | 1 | 0% | 227 | 2,004 | +783% | 0 | 0 | — |
case-06 | fail→fail | 3,840 | 5,553 | +45% | 1 | 1 | 0% | 473 | 1,351 | +186% | 0 | 0 | — |
case-07 | fail→pass | 2,439 | 6,228 | +155% | 1 | 1 | 0% | 257 | 1,416 | +451% | 0 | 0 | — |
case-08 | fail→fail | 3,285 | 9,319 | +184% | 1 | 1 | 0% | 363 | 783 | +116% | 0 | 0 | — |
case-09 | fail→pass | 8,058 | 4,535 | -44% | 1 | 1 | 0% | 1,313 | 1,311 | -0% | 0 | 0 | — |
case-10 | fail→pass | 13,893 | 3,336 | -76% | 1 | 1 | 0% | 2,185 | 1,004 | -54% | 0 | 0 | — |
case-11 | fail→fail | 2,589 | 17,183 | +564% | 1 | 1 | 0% | 364 | 1,533 | +321% | 0 | 0 | — |
case-12 | fail→fail | 6,134 | 10,141 | +65% | 1 | 1 | 0% | 1,040 | 1,131 | +9% | 0 | 0 | — |
case-13 | fail→pass | 10,594 | 2,643 | -75% | 1 | 1 | 0% | 1,779 | 882 | -50% | 0 | 0 | — |
case-14 | fail→fail | 3,153 | 21,953 | +596% | 1 | 1 | 0% | 441 | 2,776 | +529% | 0 | 0 | — |
case-15 | fail→fail | 3,983 | 12,015 | +202% | 1 | 1 | 0% | 570 | 1,235 | +117% | 0 | 0 | — |
case-16 | fail→fail | 3,514 | 12,079 | +244% | 1 | 1 | 0% | 457 | 1,890 | +314% | 0 | 0 | — |
case-17 | fail→pass | 15,232 | 4,927 | -68% | 1 | 1 | 0% | 2,454 | 812 | -67% | 0 | 0 | — |
case-18 | fail→fail | 8,739 | 2,397 | -73% | 1 | 1 | 0% | 1,472 | 735 | -50% | 0 | 0 | — |
case-19 | fail→fail | 3,891 | 11,753 | +202% | 1 | 1 | 0% | 490 | 1,348 | +175% | 0 | 0 | — |
case-20 | pass→pass | 9,057 | 4,575 | -49% | 1 | 1 | 0% | 1,413 | 1,195 | -15% | 0 | 0 | — |
case-21 | pass→pass | 2,645 | 2,049 | -23% | 1 | 1 | 0% | 345 | 755 | +119% | 0 | 0 | — |
case-22 | pass→pass | 1,951 | 1,833 | -6% | 1 | 1 | 0% | 270 | 750 | +178% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 15 counted toward the lift figure. The other 7 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +23 percentage points is the difference between those two pass rates over the 15 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.