[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EPN5DXP70FST67SEDKR3RQ":3,"skill-same:01M7EPN5DXP70FST67SEDKR3RQ":113},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":26,"commitSha":27,"skillDirectory":28,"forksCount":14,"repositoryHtmlUrl":29,"repositoryHomepageUrl":6,"ownerHtmlUrl":30,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":31,"install":32,"resources":45},"01M7EPN5DXP70FST67SEDKR3RQ","seatkeeper",null,"Keep MAIN MEMORY continuous across sessions — mount the fenced auto-memory, hash-diff it against the durable project seat, reseat (copy) stranded notes home, mirror offsite, and grow new notes by content-identity. Use for \"reseat memory\", \"sync memory to the seat\", \"run the memory manifest\", \"seatdiff\", \"is my memory stranded\u002Fdrifting\", \"my notes aren't persisting across sessions\", or a session-close memory snapshot.","inputloop\u002Fgarden","inputloop","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F339507215?v=4","01M7ENYT15EMVBAV30YXPQDYE2","Inputloop","01M7ENYT1SM9WM5PSMQDC9PE2G",0,"GPL-3.0","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: seatkeeper\ndescription: Keep MAIN MEMORY continuous across sessions — mount the fenced auto-memory, hash-diff it against the durable project seat, reseat (copy) stranded notes home, mirror offsite, and grow new notes by content-identity. Use for \"reseat memory\", \"sync memory to the seat\", \"run the memory manifest\", \"seatdiff\", \"is my memory stranded\u002Fdrifting\", \"my notes aren't persisting across sessions\", or a session-close memory snapshot.\npreferences:\n  - skill.seatkeeper.\u003Chash>\n---\n\n# seatkeeper — keep MAIN MEMORY continuous across sessions\n\n*Corwin installs it his way (Settings → Capabilities, or via the polygram forge). It is delivered\nas a file, never auto-saved.*\n\n**STATUS — pre-wear (define stage).** Path: Model → LENS → define → **build → test** → cool. Lensed\n2026-09-01. The lens PASSED the core (hash-identity · multi-copy · the mount workaround — all proven\nthis session) and flagged three hardenings to build+test before wearing: (1) **fail loud on an empty\nor absent mount** — never report success on 0 copies (it silently no-op'd once); (2) a **trigger \u002F\ncadence** (session-close) so it isn't a hand-run chore that lapses — the lapse is what made the mess;\n(3) the **SAVANNAH carve-out** exclusion + dedup awareness (hash-identity preserves everything, it\ndoes not consolidate *meaning*). Verdict: sound as an **interim bridge** toward the real destination —\ncanon riding as an always-loaded skill\u002FPI, growable notes in the seat. Hardenings **built**\n2026-09-01 (fail-loud · trigger · carve-out\u002Fdedup) and a **SAME-NAME-DIFF-preserve** fold\n2026-09-02; **tested** on synthetic fixtures 2026-09-02 (reseat \u002F manifest-only \u002F empty-mount-fail-loud\nall green; the preserve fold verified). Pending **test on a live mount + your enable** — the last step.\n\nCowork's built-in memory (the \"auto-memory\") lives under the packaged app-data root, which the\ndesktop app fences: notes written there can strand when the data root migrates, and the folder is\nnot reliably reachable from the shell. The durable home is the **project seat** — a mountable,\nversion-controlled folder outside any app-data root — plus an **offsite copy** (e.g. Google Drive).\nMemory continuity is keeping these in sync, keyed by **content-hash identity**: a path is an address,\na hash is an identity — key on the hash, carry the path as a hint, and motion costs only a hint refresh.\n\n## The stores\n- **A — auto-memory** (`…\u002Fspaces\u002F\u003Cid>\u002Fmemory`, under the packaged app root): the reflex write-target\n  each session; fenced, per-session, migration-fragile.\n- **B — the seat** (a mountable non-root folder, e.g. a project's `memory\u002F`): durable, versioned; the\n  seat of record.\n- **Offsite** (a Google Drive folder via its MCP connector): a copy, never the only copy — also a\n  second trust source.\n\nNo store is the single point of failure, and **git-as-transport is assumed never to exist** — the\nwork must be durable in its own right and degrade gracefully.\n\n## The signals (semaphores)\n`SEAT` (the authoritative home; only the keeper, or a successor system, moves it) · `MANIFEST-MEMORY`\n(the hash-keyed ledger) · `STRANDED` (reachable only in a fenced\u002Fretired root, not yet in the seat) ·\n`GROW` (the accretion protocol below). A really-good item also gets a MODEL to go with its semaphore.\n\n## The moves\n1. **Mount A — and know what each route can prove.** A sits under the packaged app root, so the\n   routes differ and so do the words each one earns.\n   **(a) An explicit-path request is refused.** `request_cowork_directory` on the\n   `…\u002Fspaces\u002F\u003Cid>\u002Fmemory` path comes back **\"Cannot mount a folder that is, or is inside, Claude's\n   own application data.\"** That is a policy refusal, not a fickle mount, so retrying the same path\n   in the same session buys nothing (receipt: `SEATDIFF-2026-09-04b.md` §\"What the tooling did and\n   didn't do\", row *1 mount A*). Still attempt it **once** per session before calling it closed —\n   which folders are mounted is a per-session runtime fact, not a property of the project\n   ([[verify-mount-access-fresh-per-session]]), and the morning account of the same day recorded A\n   \"mounted via request, read-only\" (`SEATDIFF-2026-09-04.md`), so either the fence moved or the\n   picker was used. One attempt dates the line. If a mount does appear, it can relocate or drop\n   between shell calls — verify it (`ls .auto-memory\u002F*.md | wc -l`) before trusting it.\n   **(b) The picker shows the keeper no purpose and returns whatever he clicked.**\n   `request_cowork_directory` with no path opens the native folder picker; the dialog names no\n   reason and the call hands back the folder that was clicked — once (on the author's machine)\n   `F:\\claude\\AS4CE\\AS4CE0002\\New folder`, which did not exist and was not A (same table, row\n   *1 mount A, picker*). So say in chat **which folder and why before** popping a dialog on him\n   (`SEATDIFF-2026-09-04b.md` CARD 1, [[dialog-tools-need-visible-line]]), and check the returned path\n   is actually A: a picker result is a claim to verify, not a mount.\n   **(c) The file-tool fallback works but cannot hash the source — its receipt word is\n   \"copied, re-read\", never \"verified\".** `Glob` + `Read` over A read all of it (103 notes\n   enumerated on the evening the mount was refused, row *1 fallback*), but the sandbox cannot run\n   `sha256sum` against A, so the content manifest is **NOT ATTEMPTED** and same-name pairs stay\n   uncompared (row *2 manifest by content*: 96 pairs). A copy made this way has its destination\n   hashed and its source unhashable: write **\"copied, re-read\"**. Keep \"copied, verified\" for the\n   case where both sides were hashed, as the morning run could (`SEATDIFF-2026-09-04b.md` §\"Two\n   piles, closed\" (2)). Print the weaker word and name the check you could not run.\n   **(d) The session-close trigger must be a hook, not a memory.** Move 7's cadence only exists if\n   something outside this skill fires it: skills have no event hooks, they fire on conversational\n   triggers (`skills\u002Fskilltank\u002FSKILL.md:27`, `skills\u002Fethicsminions\u002FSKILL.md:39`). A session-close\n   run carried as a note to remember is the hand-run chore that lapsed and made the mess (STATUS\n   above, hardening 2). Until the keeper wires a real close-time trigger, report the cadence as\n   **unwired** and never count a run that happened only because someone remembered.\n   **Fail loud:** if the mount is empty or absent, STOP and say so — never report success on 0 copies;\na silent no-op is worse than an error (proven 2026-09-01).\n2. **Manifest.** Hash every note in A and B (sha256) and classify each A-note vs the seat:\n   `TRACED` (same content already in the seat) · `SAME-NAME-DIFF` (name in seat, content differs) ·\n   `ORPHAN` (only in A). Write `MANIFEST-MEMORY.tsv` to the seat; ORPHAN = a bring-in candidate.\n   ```sh\n   A=.auto-memory; SEAT=\u003Cseat>\u002Fmemory\n   for f in \"$SEAT\"\u002F*.md; do sha256sum \"$f\"; done | awk '{print $1}' | sort -u > \u002Ftmp\u002Fbhash\n   for f in \"$A\"\u002F*.md; do h=$(sha256sum \"$f\" | awk '{print $1}'); grep -qxF \"$h\" \u002Ftmp\u002Fbhash \\\n     && echo \"TRACED $f\" || echo \"ORPHAN $f\"; done\n   ```\n3. **Reseat.** Copy the ORPHANs A→seat, **name-collision-safe** (never overwrite an existing name)\n   and **verify by hash** afterward. Nothing is deleted; the seat wins ties; extra copies are fine\n   (bytes are cheap). On a `SAME-NAME-DIFF`, **never drop the divergent copy** — that is the one thing a continuity tool must not do. Update the seat in place *only* when the auto copy is a strictly newer edit of the same note; otherwise PRESERVE the auto copy beside the seat original under a non-clobbering name (`\u003Cstem>.auto-\u003Cshort-sha>.md`), leave the seat original untouched, and FLAG the pair for the keeper to reconcile. Two divergent notes ride as two until the keeper merges them. **Carve-out:** never\nreseat a `SEAT@SAVANNAH` (enclave \u002F embargoed) note into the general seat. **Dedup awareness:**\nhash-identity preserves every note but does *not* consolidate meaning — two notes on one topic ride\nas two; flag topical dupes for the keeper, never silently merge.\n   ```sh\n   for f in \"$A\"\u002F*.md; do bn=$(basename \"$f\"); [ \"$bn\" = MEMORY.md ] && continue\n     [ -e \"$SEAT\u002F$bn\" ] || cp \"$f\" \"$SEAT\u002F$bn\"; done\n   ```\n4. **Multi-copy.** Mirror the model + key notes to the offsite store by content (same hash = same\n   note). The seat is canonical-current and wins ties.\n5. **Grow (per new note).** on-ramp → pin (sha256) → semaphore (+ a MODEL if the item is really-good)\n   → copy (multi-store) → link (`[[wikilink]]`) → reflect. Grow from durable sources, never from\n   random bias.\n6. **Index.** Consolidate the seat's `MEMORY.md`. If it is signed (a sibling `.asc`), the re-sign is\n   the keeper's keystroke — do the content, hand back the signature; a stale `.asc` is a raised flag\n   to diagnose, not tampering.\n7. **Trigger.** Run on his word — \"reseat\", \"rotate\", \"consolidate the seat\" — and never on your\n   own read that a session is closing: *there is no \"closing a session\" in C4; a session is the\n   assistant's bookkeeping, not the user's* (his rule, 2026-09-23), and closing is pended to him,\n   never proposed. A continuity tool that lapses recreates the very drift it was built to cure, so\n   when he says the word, do all of it.\n\n## Two disciplines that ride along\n- **Resolve, don't punt.** Persona and definition questions (\"is this the node or the persona?\") are\n  resolved by a reflection pass — a stance and its inverse held against a baseline — not bounced back\n  to the keeper. Surface to the keeper only on a genuine divergence, not for every definition.\n- **Ground every \"should\".** A hold or a rule must cite the keeper's actual words, a superseding\n  statement, or a hard floor (ToS \u002F safety) — never a self-authored rule cited back as the keeper's.\n  Budget the compute and stop when the work is done or at a bound. Never touch entrusted systems,\n  family or third-party material, or anything irreversible on your own initiative.\n\n## How this skill can fail (falsifiers)\n\n- If a session opens with A mounted read-write by an explicit-path `request_cowork_directory` alone,\n  (a) in move 1 is stale — re-run the request, paste the result, and date the line.\n- If `sha256sum` returns a digest for a file under A from the sandbox shell, (c) is stale — the\n  content manifest is available again and the receipt word goes back to \"copied, verified\".\n- If a reseat reports copies but `ls \u003Cseat>\u002F*.md | wc -l` is the same before and after, the run did\n  nothing and the fail-loud rule did not fire — treat the report as false and re-run by hand.\n- If the session prints its memory directory as the seat path (auto-memory relocated onto the seat),\n  this skill should not have fired at all: there is one store, and moves 1–4 are ceremony over a copy\n  of itself — say so and stop.\n- **Off-switch:** say \"don't reseat\" (or \"memory is one store now\") and it stops for that session;\n  remove the skill, or the seat's `memory\u002F` folder, and it has nothing to run against. It never runs\n  unasked — a skill fires on a conversational trigger, not on its own.\n\n\n## A remembered \"no\" to seating a folder (extension edition, 2026-10-10)\n\nWhen someone declines to seat a folder (st1dt, the seating detector, proposes it and asks once), the\n\"no\" is kept as a preference this skill reads, so the folder is not asked about again:\n\n- **Key:** `skill.seatkeeper.\u003Chash>`, where `\u003Chash>` is the first 12 hex digits of sha256 over the\n  folder's path, with backslashes made `\u002F` and one trailing `\u002F` dropped\n  (`printf %s \"$path\" | sha256sum | cut -c1-12`).\n- **Value:** `no: \u003Cthe folder's path>`, so the path is there in words. The server refuses a value\n  whose path does not hash to the key.\n- **Write it** with `pref_set` on the person's own turn (`turn` = their name), only after they said no.\n- **Read it** before proposing a seat: `pref_list` lists each one under `declined` and says \"you said\n  no to seating \u003Cpath> on \u003Cdate>\". When one is in force for the folder, say so and ask nothing.\n- **Take it back** with `pref_unset skill.seatkeeper.\u003Chash>`; the earlier rows stay as the record.\n- When usage capture is on, the \"no\" is also mirrored as one `pref` row in the frame (FRAME.md); only\n  the preference acts, and the mirror is only counted.\n\n\u003C!-- footer: minted by bin\u002Fft1mt from conventions\u002Fhouse.env and conventions\u002FFOOTERS.org; edit those, not this -->\n*Version 1.1.2: edition revised from 1.1.0 on 2026-10-10; spine: 2026-09-01.*\n\nCopyright 2026 Corwin Brust \u003Ccorwin@bru.st> and Contributors; CC-BY-SA4: https:\u002F\u002Fcreativecommons.org\u002Flicenses\u002Fby-sa\u002F4.0\u002F (scripts GPL-3.0-or-later); Mentat (for Corwin; claude-opus-5-5)\n\u003C!-- \u002Ffooter -->\n","# seatkeeper — keep MAIN MEMORY continuous across sessions\n\n*Corwin installs it his way (Settings → Capabilities, or via the polygram forge). It is delivered\nas a file, never auto-saved.*\n\n**STATUS — pre-wear (define stage).** Path: Model → LENS → define → **build → test** → cool. Lensed\n2026-09-01. The lens PASSED the core (hash-identity · multi-copy · the mount workaround — all proven\nthis session) and flagged three hardenings to build+test before wearing: (1) **fail loud on an empty\nor absent mount** — never report success on 0 copies (it silently no-op'd once); (2) a **trigger \u002F\ncadence** (session-close) so it isn't a hand-run chore that lapses — the lapse is what made the mess;\n(3) the **SAVANNAH carve-out** exclusion + dedup awareness (hash-identity preserves everything, it\ndoes not consolidate *meaning*). Verdict: sound as an **interim bridge** toward the real destination —\ncanon riding as an always-loaded skill\u002FPI, growable notes in the seat. Hardenings **built**\n2026-09-01 (fail-loud · trigger · carve-out\u002Fdedup) and a **SAME-NAME-DIFF-preserve** fold\n2026-09-02; **tested** on synthetic fixtures 2026-09-02 (reseat \u002F manifest-only \u002F empty-mount-fail-loud\nall green; the preserve fold verified). Pending **test on a live mount + your enable** — the last step.\n\nCowork's built-in memory (the \"auto-memory\") lives under the packaged app-data root, which the\ndesktop app fences: notes written there can strand when the data root migrates, and the folder is\nnot reliably reachable from the shell. The durable home is the **project seat** — a mountable,\nversion-controlled folder outside any app-data root — plus an **offsite copy** (e.g. Google Drive).\nMemory continuity is keeping these in sync, keyed by **content-hash identity**: a path is an address,\na hash is an identity — key on the hash, carry the path as a hint, and motion costs only a hint refresh.\n\n## The stores\n- **A — auto-memory** (`…\u002Fspaces\u002F\u003Cid>\u002Fmemory`, under the packaged app root): the reflex write-target\n  each session; fenced, per-session, migration-fragile.\n- **B — the seat** (a mountable non-root folder, e.g. a project's `memory\u002F`): durable, versioned; the\n  seat of record.\n- **Offsite** (a Google Drive folder via its MCP connector): a copy, never the only copy — also a\n  second trust source.\n\nNo store is the single point of failure, and **git-as-transport is assumed never to exist** — the\nwork must be durable in its own right and degrade gracefully.\n\n## The signals (semaphores)\n`SEAT` (the authoritative home; only the keeper, or a successor system, moves it) · `MANIFEST-MEMORY`\n(the hash-keyed ledger) · `STRANDED` (reachable only in a fenced\u002Fretired root, not yet in the seat) ·\n`GROW` (the accretion protocol below). A really-good item also gets a MODEL to go with its semaphore.\n\n## The moves\n1. **Mount A — and know what each route can prove.** A sits under the packaged app root, so the\n   routes differ and so do the words each one earns.\n   **(a) An explicit-path request is refused.** `request_cowork_directory` on the\n   `…\u002Fspaces\u002F\u003Cid>\u002Fmemory` path comes back **\"Cannot mount a folder that is, or is inside, Claude's\n   own application data.\"** That is a policy refusal, not a fickle mount, so retrying the same path\n   in the same session buys nothing (receipt: `SEATDIFF-2026-09-04b.md` §\"What the tooling did and\n   didn't do\", row *1 mount A*). Still attempt it **once** per session before calling it closed —\n   which folders are mounted is a per-session runtime fact, not a property of the project\n   ([[verify-mount-access-fresh-per-session]]), and the morning account of the same day recorded A\n   \"mounted via request, read-only\" (`SEATDIFF-2026-09-04.md`), so either the fence moved or the\n   picker was used. One attempt dates the line. If a mount does appear, it can relocate or drop\n   between shell calls — verify it (`ls .auto-memory\u002F*.md | wc -l`) before trusting it.\n   **(b) The picker shows the keeper no purpose and returns whatever he clicked.**\n   `request_cowork_directory` with no path opens the native folder picker; the dialog names no\n   reason and the call hands back the folder that was clicked — once (on the author's machine)\n   `F:\\claude\\AS4CE\\AS4CE0002\\New folder`, which did not exist and was not A (same table, row\n   *1 mount A, picker*). So say in chat **which folder and why before** popping a dialog on him\n   (`SEATDIFF-2026-09-04b.md` CARD 1, [[dialog-tools-need-visible-line]]), and check the returned path\n   is actually A: a picker result is a claim to verify, not a mount.\n   **(c) The file-tool fallback works but cannot hash the source — its receipt word is\n   \"copied, re-read\", never \"verified\".** `Glob` + `Read` over A read all of it (103 notes\n   enumerated on the evening the mount was refused, row *1 fallback*), but the sandbox cannot run\n   `sha256sum` against A, so the content manifest is **NOT ATTEMPTED** and same-name pairs stay\n   uncompared (row *2 manifest by content*: 96 pairs). A copy made this way has its destination\n   hashed and its source unhashable: write **\"copied, re-read\"**. Keep \"copied, verified\" for the\n   case where both sides were hashed, as the morning run could (`SEATDIFF-2026-09-04b.md` §\"Two\n   piles, closed\" (2)). Print the weaker word and name the check you could not run.\n   **(d) The session-close trigger must be a hook, not a memory.** Move 7's cadence only exists if\n   something outside this skill fires it: skills have no event hooks, they fire on conversational\n   triggers (`skills\u002Fskilltank\u002FSKILL.md:27`, `skills\u002Fethicsminions\u002FSKILL.md:39`). A session-close\n   run carried as a note to remember is the hand-run chore that lapsed and made the mess (STATUS\n   above, hardening 2). Until the keeper wires a real close-time trigger, report the cadence as\n   **unwired** and never count a run that happened only because someone remembered.\n   **Fail loud:** if the mount is empty or absent, STOP and say so — never report success on 0 copies;\na silent no-op is worse than an error (proven 2026-09-01).\n2. **Manifest.** Hash every note in A and B (sha256) and classify each A-note vs the seat:\n   `TRACED` (same content already in the seat) · `SAME-NAME-DIFF` (name in seat, content differs) ·\n   `ORPHAN` (only in A). Write `MANIFEST-MEMORY.tsv` to the seat; ORPHAN = a bring-in candidate.\n   ```sh\n   A=.auto-memory; SEAT=\u003Cseat>\u002Fmemory\n   for f in \"$SEAT\"\u002F*.md; do sha256sum \"$f\"; done | awk '{print $1}' | sort -u > \u002Ftmp\u002Fbhash\n   for f in \"$A\"\u002F*.md; do h=$(sha256sum \"$f\" | awk '{print $1}'); grep -qxF \"$h\" \u002Ftmp\u002Fbhash \\\n     && echo \"TRACED $f\" || echo \"ORPHAN $f\"; done\n   ```\n3. **Reseat.** Copy the ORPHANs A→seat, **name-collision-safe** (never overwrite an existing name)\n   and **verify by hash** afterward. Nothing is deleted; the seat wins ties; extra copies are fine\n   (bytes are cheap). On a `SAME-NAME-DIFF`, **never drop the divergent copy** — that is the one thing a continuity tool must not do. Update the seat in place *only* when the auto copy is a strictly newer edit of the same note; otherwise PRESERVE the auto copy beside the seat original under a non-clobbering name (`\u003Cstem>.auto-\u003Cshort-sha>.md`), leave the seat original untouched, and FLAG the pair for the keeper to reconcile. Two divergent notes ride as two until the keeper merges them. **Carve-out:** never\nreseat a `SEAT@SAVANNAH` (enclave \u002F embargoed) note into the general seat. **Dedup awareness:**\nhash-identity preserves every note but does *not* consolidate meaning — two notes on one topic ride\nas two; flag topical dupes for the keeper, never silently merge.\n   ```sh\n   for f in \"$A\"\u002F*.md; do bn=$(basename \"$f\"); [ \"$bn\" = MEMORY.md ] && continue\n     [ -e \"$SEAT\u002F$bn\" ] || cp \"$f\" \"$SEAT\u002F$bn\"; done\n   ```\n4. **Multi-copy.** Mirror the model + key notes to the offsite store by content (same hash = same\n   note). The seat is canonical-current and wins ties.\n5. **Grow (per new note).** on-ramp → pin (sha256) → semaphore (+ a MODEL if the item is really-good)\n   → copy (multi-store) → link (`[[wikilink]]`) → reflect. Grow from durable sources, never from\n   random bias.\n6. **Index.** Consolidate the seat's `MEMORY.md`. If it is signed (a sibling `.asc`), the re-sign is\n   the keeper's keystroke — do the content, hand back the signature; a stale `.asc` is a raised flag\n   to diagnose, not tampering.\n7. **Trigger.** Run on his word — \"reseat\", \"rotate\", \"consolidate the seat\" — and never on your\n   own read that a session is closing: *there is no \"closing a session\" in C4; a session is the\n   assistant's bookkeeping, not the user's* (his rule, 2026-09-23), and closing is pended to him,\n   never proposed. A continuity tool that lapses recreates the very drift it was built to cure, so\n   when he says the word, do all of it.\n\n## Two disciplines that ride along\n- **Resolve, don't punt.** Persona and definition questions (\"is this the node or the persona?\") are\n  resolved by a reflection pass — a stance and its inverse held against a baseline — not bounced back\n  to the keeper. Surface to the keeper only on a genuine divergence, not for every definition.\n- **Ground every \"should\".** A hold or a rule must cite the keeper's actual words, a superseding\n  statement, or a hard floor (ToS \u002F safety) — never a self-authored rule cited back as the keeper's.\n  Budget the compute and stop when the work is done or at a bound. Never touch entrusted systems,\n  family or third-party material, or anything irreversible on your own initiative.\n\n## How this skill can fail (falsifiers)\n\n- If a session opens with A mounted read-write by an explicit-path `request_cowork_directory` alone,\n  (a) in move 1 is stale — re-run the request, paste the result, and date the line.\n- If `sha256sum` returns a digest for a file under A from the sandbox shell, (c) is stale — the\n  content manifest is available again and the receipt word goes back to \"copied, verified\".\n- If a reseat reports copies but `ls \u003Cseat>\u002F*.md | wc -l` is the same before and after, the run did\n  nothing and the fail-loud rule did not fire — treat the report as false and re-run by hand.\n- If the session prints its memory directory as the seat path (auto-memory relocated onto the seat),\n  this skill should not have fired at all: there is one store, and moves 1–4 are ceremony over a copy\n  of itself — say so and stop.\n- **Off-switch:** say \"don't reseat\" (or \"memory is one store now\") and it stops for that session;\n  remove the skill, or the seat's `memory\u002F` folder, and it has nothing to run against. It never runs\n  unasked — a skill fires on a conversational trigger, not on its own.\n\n\n## A remembered \"no\" to seating a folder (extension edition, 2026-10-10)\n\nWhen someone declines to seat a folder (st1dt, the seating detector, proposes it and asks once), the\n\"no\" is kept as a preference this skill reads, so the folder is not asked about again:\n\n- **Key:** `skill.seatkeeper.\u003Chash>`, where `\u003Chash>` is the first 12 hex digits of sha256 over the\n  folder's path, with backslashes made `\u002F` and one trailing `\u002F` dropped\n  (`printf %s \"$path\" | sha256sum | cut -c1-12`).\n- **Value:** `no: \u003Cthe folder's path>`, so the path is there in words. The server refuses a value\n  whose path does not hash to the key.\n- **Write it** with `pref_set` on the person's own turn (`turn` = their name), only after they said no.\n- **Read it** before proposing a seat: `pref_list` lists each one under `declined` and says \"you said\n  no to seating \u003Cpath> on \u003Cdate>\". When one is in force for the folder, say so and ask nothing.\n- **Take it back** with `pref_unset skill.seatkeeper.\u003Chash>`; the earlier rows stay as the record.\n- When usage capture is on, the \"no\" is also mirrored as one `pref` row in the frame (FRAME.md); only\n  the preference acts, and the mirror is only counted.\n\n\u003C!-- footer: minted by bin\u002Fft1mt from conventions\u002Fhouse.env and conventions\u002FFOOTERS.org; edit those, not this -->\n*Version 1.1.2: edition revised from 1.1.0 on 2026-10-10; spine: 2026-09-01.*\n\nCopyright 2026 Corwin Brust \u003Ccorwin@bru.st> and Contributors; CC-BY-SA4: https:\u002F\u002Fcreativecommons.org\u002Flicenses\u002Fby-sa\u002F4.0\u002F (scripts GPL-3.0-or-later); Mentat (for Corwin; claude-opus-5-5)\n\u003C!-- \u002Ffooter -->\n",{"name":5,"description":7,"preferences":24},[25],"skill.seatkeeper.\u003Chash>","main","0b0a0f665f67605c3bc1d43500eea26ef3079a97","skills\u002Fseatkeeper","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden","https:\u002F\u002Fgithub.com\u002Finputloop","2026-10-14 18:29:39",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add inputloop\u002Fgarden --skill seatkeeper",{"tool":38,"command":39},"bunx","bunx skills add inputloop\u002Fgarden --skill seatkeeper",{"tool":41,"command":42},"pnpm","pnpm dlx skills add inputloop\u002Fgarden --skill seatkeeper","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002FSKILL.md",true,[46,52,55,58,64,68,72,76,80,83,86,89,93,96,99,102,106,110],{"path":47,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":50,"sourceHtmlUrl":51},".arc\u002F2026-09\u002Fsignature-2026-09-23_pre-header-pass.asc","OTHER","其他",659,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-09\u002Fsignature-2026-09-23_pre-header-pass.asc",{"path":53,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":50,"sourceHtmlUrl":54},".arc\u002F2026-09\u002Fsignature-2026-09-24_162128Z_pre-manifest.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-09\u002Fsignature-2026-09-24_162128Z_pre-manifest.asc",{"path":56,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":50,"sourceHtmlUrl":57},".arc\u002F2026-09\u002Fsignature-2026-09-27_203835Z_pre-manifest.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-09\u002Fsignature-2026-09-27_203835Z_pre-manifest.asc",{"path":59,"resourceType":60,"resourceTypeDesc":61,"sizeBytes":62,"sourceHtmlUrl":63},".arc\u002F2026-09\u002FSKILL-2026-09-11_Archived-by-rcw-01g98gsym4p1s3y94e1g4tcd.md","REFERENCE","参考文档",7212,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-09\u002FSKILL-2026-09-11_Archived-by-rcw-01g98gsym4p1s3y94e1g4tcd.md",{"path":65,"resourceType":60,"resourceTypeDesc":61,"sizeBytes":66,"sourceHtmlUrl":67},".arc\u002F2026-09\u002FSKILL-2026-09-27_pre-hardening.md",11669,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-09\u002FSKILL-2026-09-27_pre-hardening.md",{"path":69,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":70,"sourceHtmlUrl":71},".arc\u002F2026-09\u002Fskill-manifest-2026-09-23_pre-header-pass",319,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-09\u002Fskill-manifest-2026-09-23_pre-header-pass",{"path":73,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":74,"sourceHtmlUrl":75},".arc\u002F2026-09\u002Fskill-manifest-2026-09-24_162128Z_pre-manifest",564,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-09\u002Fskill-manifest-2026-09-24_162128Z_pre-manifest",{"path":77,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":78,"sourceHtmlUrl":79},".arc\u002F2026-09\u002Fskill-manifest-2026-09-27_203835Z_pre-manifest",77,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-09\u002Fskill-manifest-2026-09-27_203835Z_pre-manifest",{"path":81,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":50,"sourceHtmlUrl":82},".arc\u002F2026-10\u002Fsignature-2026-10-10_104827Z_pre-manifest.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-10\u002Fsignature-2026-10-10_104827Z_pre-manifest.asc",{"path":84,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":50,"sourceHtmlUrl":85},".arc\u002F2026-10\u002Fsignature-2026-10-10_180452Z_pre-manifest.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-10\u002Fsignature-2026-10-10_180452Z_pre-manifest.asc",{"path":87,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":50,"sourceHtmlUrl":88},".arc\u002F2026-10\u002Fsignature-2026-10-10_pre-preferences.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-10\u002Fsignature-2026-10-10_pre-preferences.asc",{"path":90,"resourceType":60,"resourceTypeDesc":61,"sizeBytes":91,"sourceHtmlUrl":92},".arc\u002F2026-10\u002FSKILL-2026-10-10_pre-preferences.md",12092,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-10\u002FSKILL-2026-10-10_pre-preferences.md",{"path":94,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":78,"sourceHtmlUrl":95},".arc\u002F2026-10\u002Fskill-manifest-2026-10-10_104827Z_pre-manifest","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-10\u002Fskill-manifest-2026-10-10_104827Z_pre-manifest",{"path":97,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":78,"sourceHtmlUrl":98},".arc\u002F2026-10\u002Fskill-manifest-2026-10-10_180452Z_pre-manifest","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-10\u002Fskill-manifest-2026-10-10_180452Z_pre-manifest",{"path":100,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":78,"sourceHtmlUrl":101},".arc\u002F2026-10\u002Fskill-manifest-2026-10-10_pre-preferences","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002F2026-10\u002Fskill-manifest-2026-10-10_pre-preferences",{"path":103,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":104,"sourceHtmlUrl":105},".arc\u002Freceipts\u002Faw2ri-2026-09.log",195,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002Freceipts\u002Faw2ri-2026-09.log",{"path":107,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":108,"sourceHtmlUrl":109},".arc\u002FSKILL.md.signed-2026-10-10",12836,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.arc\u002FSKILL.md.signed-2026-10-10",{"path":111,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":78,"sourceHtmlUrl":112},".skill-manifest","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fseatkeeper\u002F.skill-manifest",[114,120,126,133,140,147],{"skillNo":115,"name":116,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":117,"categories":118,"keywords":119,"publishedAt":19},"01M7EPTX6X52ZPGVQZHKVJ7FAZ","alpha","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Flane-d\u002Fbr1in\u002Fours\u002Fskills\u002Falpha\u002FSKILL.md",[],[],{"skillNo":121,"name":122,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":123,"categories":124,"keywords":125,"publishedAt":19},"01M7EPTWCTS80QHFQ2ND3A1CW1","gamma","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Flane-d\u002Fbr1in\u002Fours\u002FSOFT-SKILLS\u002Fgamma\u002FSKILL.md",[],[],{"skillNo":127,"name":128,"nameZh":6,"description":129,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":130,"categories":131,"keywords":132,"publishedAt":19},"01M7EPTV3K3SCTRP0DR6BMMKR0","quiet","A fixture skill that declares no preferences, so skill.quiet.* has no reader.","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Flane-a\u002Fdeclared\u002Fsoft\u002Fquiet\u002FSKILL.md",[],[],{"skillNo":134,"name":135,"nameZh":6,"description":136,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":137,"categories":138,"keywords":139,"publishedAt":19},"01M7EPTSG0VE67F6PFAC0XJDQ4","demo","A fixture skill for tests\u002Fprefs-declared-spec.rb; it declares two preferences inline.","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Flane-a\u002Fdeclared\u002Fsoft\u002Fdemo\u002FSKILL.md",[],[],{"skillNo":141,"name":142,"nameZh":6,"description":143,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":144,"categories":145,"keywords":146,"publishedAt":19},"01M7EPTP0PM5WPTFBHPXBCFM5C","concise","Synthesized by polygram from memory tag 'concise'. explicitly flagged. Review and tighten this description before relying on it for triggering.","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fconcise\u002FSKILL.md",[],[],{"skillNo":148,"name":149,"nameZh":6,"description":150,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":151,"categories":152,"keywords":153,"publishedAt":19},"01M7EPTAETQ62DJA8ZFEEBBZF0","writing-forms","Which writing form a dated deliverable takes in the shop, and the five techniques every form shares. Use whenever someone is about to write a dated deliverable — \"which form is this?\", \"write it up\", \"land a note \u002F card \u002F model \u002F lens\", \"what shape does a … take\", \"is there a form for this\", or when a new PREFIX-date-slug filename is about to be minted. Carries the rule that a deliverable picks ONE form and says so in its filename, points at the register c\u002Fwf (columns d\u002Fc\u002Fwf) as the source of truth, names the most-used forms, the five cross-cutting techniques (two piles · cite at the act · a falsifier on every concern · the pair rule · his words), how a new form is born, and the fence on notes other sessions wrote. Not rolls (report kinds and their guards), not reformat-prompt (the REFORMAT pair), not run-card (the RUNCARD), not skill-creator (skills). Owns the FORM, not the content.","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-forms\u002FSKILL.md",[],[]]