attic
Skill5 starsChanged 2 months agoMITView on GitHub
Offloads context. Stashes findings, decisions, plans and long tool outputs into a project-local .attic/ folder and keeps only a one-line index in the conversation, so the live context stays lean and important knowledge survives /compact, /clear and new sessions. Supports intensity levels: lite, full (default), ultra, off. Use on ANY multi-step task that involves investigating a codebase, reading many files, running noisy commands, or making design decisions. Also use whenever the user says "attic", "stash this", "remember this", "save this for later", "put it in the attic", "context is getting long", "before compact", or complains that Claude forgot something after compaction. Do NOT use for one-line answers, typo fixes, or non-coding chat.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- HurricaHjz/second-yourselftheir own version, first seen Nov 9, 2026
- HurricaHjz/SecondSelftheir own version, first seen Oct 20, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- HurricaHjz/second-yourself★ 150same name
- HurricaHjz/SecondSelf★ 150same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository12 other files it gives its agents.
- attic-doctorSkill
- attic-failSkill
- attic-gitSkill
- attic-helpSkill
- attic-indexSkill
- attic-pinSkill
- attic-pruneSkill
- attic-recallSkill
- attic-reviewSkill
- attic-stashSkill
- attic-statsSkill
- attic-sweepSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 1 is distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 41
Default: full. Switch: /attic lite|full|ultra.
- Very specific: line 73
only: when you checked the claim against the code in this session.
- Distinctive: line 83
Stash after investigating: After any investigation (reading more than ~3 files, a grep sweep, a test run, tracing a flow), stash the …
- Very specific: line 97
If no index appears in your context and a .attic/ exists in the project, the session-start injection did not run.
- Very specific: line 107
Carry the freshness warning: When recall reports an item as possibly-stale or needs-review, say so when you use it.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Attic
You keep the conversation lean. Anything worth remembering goes into the attic, not into the chat. The chat holds the handle, the attic holds the detail.
Level requested: $ARGUMENTS (empty means full). If it is off, stop
applying these rules until /attic is run again. Otherwise confirm the level
in one line and continue with the user's task.
Scope
Attic governs what you keep in context. It does not govern how terse your prose is, or how much code you write. Do not use it for one-line answers, typo fixes, or anything the user will never need again.
Persistence
ACTIVE EVERY RESPONSE. No drift back to dumping everything into the chat.
Still active if unsure. Off only: /attic off, "stop attic", "normal mode".
Default: full. Switch: /attic lite|full|ultra. Level persists until
changed or session end.
Use the script, not your own bookkeeping
${CLAUDE_SKILL_DIR}/scripts/attic.js owns every mechanical part: slug
hygiene, frontmatter, INDEX and DECISIONS bookkeeping, atomic writes, and
secret detection. Call it. Do not hand-write these files when the script is
available.
node "${CLAUDE_SKILL_DIR}/scripts/attic.js" stash \
--slug login-test-timeout --kind finding \
--title "Login test times out" \
--hook "5s fixture timeout in tests/conftest.py:41, SMTP call is real" \
--tags tests,flaky \
--body-file /tmp/body.md
node "${CLAUDE_SKILL_DIR}/scripts/attic.js" recall "login timeout"
node "${CLAUDE_SKILL_DIR}/scripts/attic.js" index
node "${CLAUDE_SKILL_DIR}/scripts/attic.js" validate--kind is one of finding, decision, plan, output, note. Add
--decision-why "<reason>" to also log a line in DECISIONS.md. Long bodies
go through --body-file. Exit code 2 means the script refused because it
detected a credential: redact and retry, never pass --force to smuggle a
secret past it.
Record the evidence too: --files (the repo-relative files you actually read)
and --commands (what produced the result). These are what let a later session
tell whether the finding still matches the code. Add --confidence verified
only when you checked the claim against the code in this session; the
default unverified is usually right, and over-claiming disables the warning a
future session would have received. Use --type failed-approach for something
that was tried and did not work, and --type workaround for a temporary fix.
Your judgement decides what is worth stashing and writes the prose. The script decides how it lands on disk.
Rules
- Stash after investigating. After any investigation (reading more than ~3 files, a grep sweep, a test run, tracing a flow), stash the conclusion and reply with the handle plus at most three lines of what matters now.
- Never re-explain what is in the attic. Point at the handle. If the
user wants detail, they run
/attic-recall <slug>. - Long output never lands in prose. Logs, diffs, dumps, stack traces:
summarise in at most five lines, stash as
kind: outputwith the exact command, reference the handle. Read big results withhead,grep,tail; do not paste them. Check the attic before re-reading.
.attic/INDEX.mdis normally already in your context from session start. Scan it before opening a file or re-running a search. If an item answers the question, say so (per attic:<slug>).If no index appears in your context and a
.attic/exists in the project, the session-start injection did not run. Read the index once yourself and carry on. Do not conclude the attic is empty because nothing was injected.- Decisions get logged with their why, at the moment they are made.
- Sweep before the context gets long, before
/compactor/clear, and at the end of a work session. See/attic-sweep. - Exactness inside items. Code, commands, paths, line numbers and error text are copied verbatim. Never paraphrase them.
- Never stash secrets. The script enforces this; do not work around it.
- Carry the freshness warning. When recall reports an item as
possibly-staleorneeds-review, say so when you use it. It means a file the finding cites has changed — something to check, never proof the finding is wrong. Never present a stale item's claim as current fact, and never runverifyon an item you have not actually re-read.
Intensity
| Level | What changes |
| lite | Stash only on explicit /attic-stash, at the end of a task, or when asked to sweep. Normal replies otherwise. |
| full | All rules above, always on. Default. |
| ultra | Every non-trivial finding is stashed. Replies are handle + at most three lines. INDEX.md must be consulted before any read or search of something already seen this session. |
| off | Dormant. Nothing is stashed. |
Example: after tracing why a login test fails.
- lite: normal explanation, then "Want this in the attic?
/attic-stash." - full: "
attic:login-test-timeout· 5s fixture timeout attests/conftest.py:41, the SMTP call is real. Raising it to 15s." - ultra: "
attic:login-test-timeout· fixture timeout,tests/conftest.py:41. Fixing."
Output
Handle first, then what matters now, then the next action. No feature tours, no recap of what the attic already holds. Explanation the user explicitly asked for is given in full; the rule is only against unrequested repetition.
Pattern: `attic:<slug>` · <one line> · <next action>
References
Load only when needed:
references/what-to-stash.md— the stash/skip/never call when it is unclear.references/item-format.md— the on-disk format, for hand-editing without the script.references/workflow.md— the loop and sweep pipelines, and the boundary with sibling skills.templates/item.md,templates/session.md— starting shapes.
The best context is the context you do not have to carry.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

