Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add mubit-ai/claude-plugins/plugin install mubit-memoryWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mubit-ai/claude-plugins/forget)<a href="https://agentmods.dev/skills/mubit-ai/claude-plugins/forget"><img src="https://agentmods.dev/badge/skills/mubit-ai/claude-plugins/forget/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/mubit-ai/claude-plugins/forget"><img src="https://agentmods.dev/badge/skills/mubit-ai/claude-plugins/forget.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00026 | $0.00804 |
| Opus 5.5 | $0.00010 | $0.00322 |
| Sonnet 5.5 | $0.00005 | $0.00161 |
| Haiku 4.5 | $0.00003 | $0.00080 |
Grade A, and why
forget scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Delete with the bundled script, naming the lesson — the reference_id cited in recalled
context, or the id on a line of /mubit-memory:reflect or /mubit-memory:strategies:
node ${CLAUDE_PLUGIN_ROOT}/bin/admin.mjs forget --data-dir "${MUBIT_CC_DATA_DIR:-${CLAUDE_PLUGIN_DATA}}" <lesson_id>
--data-dir "${MUBIT_CC_DATA_DIR:-${CLAUDE_PLUGIN_DATA}}" is not optional. A Bash tool
call does not inherit CLAUDE_PLUGIN_DATA — it arrives empty — so without the flag the script
has to guess which of several mubit-memory* directories the host is using, and either
refuses with no_run or acts on a run this session's hooks never touch. The host substitutes
${CLAUDE_PLUGIN_DATA} into this file before you read it, and the shell default around it
lets a session that pins MUBIT_CC_DATA_DIR keep its pin — the same order every hook resolves
in. Pass the whole expression straight through. Do not turn it into an ENV=… node … prefix:
that is no longer a node command and will stop for a permission prompt.
Confirm the id and the text with the user before you run it; there is no dry run. The script
deletes one lesson by POST /v2/control/lessons/delete and nothing else: it has no way to
delete a whole run, on purpose, because that operation removed every capture, trace and
lesson in the run and was one argument away from a single-lesson delete.
There is also no "mark superseded" operation: nothing flags an entry as replaced. Superseding is done the way described below — write the corrected lesson, and down-weight the old one so retrieval stops preferring it.
Read this before deleting
Deletion is not undoable. There is no tombstone to restore from, no recycle bin, and the entry's accumulated reinforcement history — every outcome that ever credited or blamed it — goes with it.
A lesson that is merely wrong is usually better handled with mubit_outcome and a negative
signal:
mubit_outcome reference_id=<the lesson> outcome=failure signal=-1
rationale="<why it was wrong, in one sentence>"
The promotion pipeline acts on that. The lesson's confidence drops, it stops surfacing near the top of recall, and — this is the part deletion cannot do — the reason it was wrong is now part of the record. A deleted lesson teaches the system nothing. A down-weighted one teaches it something, and the correction survives to shape what gets promoted next.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago Changed · +13 lines · -8 tokens per session a4c0c70fcfe3
- 2mo ago First seen · 49 lines · 34 tokens per session scan A 99986c81e5fa
forget is a skill published in the GitHub repository mubit-ai/claude-plugins (12 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 26 tokens to every session and 804 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
forget compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| forgetmubit-ai/claude-plugins | Skill | 12 | 1mo ago | Apache-2.0 | A | 26 | Claude Code |
| knowledge-ingestopenclaw/openclaw | Skill | 394k | yesterday | MIT | A | 80 | Claude Code, Codex |
| mem0-oss-to-platformmem0ai/mem0 | Skill | 68k | 2d ago | Apache-2.0 | A | 217 | Claude Code, Codex |
| forgetasgeirtj/system_prompts_leaks | Skill | 71k | 2d ago | CC0-1.0 | A | 74 | Claude Code, Codex |
| Cortexdanielmiessler/LifeOS | Skill | 20k | yesterday | MIT | A | 184 | Claude Code |
| memoryopensquilla/opensquilla | Skill | 7.1k | 6d ago | Apache-2.0 | A | 44 | Claude Code, Codex |
| ha-data-storesshiwenwen/hope-agent | Skill | 1.9k | yesterday | MIT | A | 115 | Claude Code, Codex |