[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EPAX3GX74QX4FSRGFYP405":3,"skill-same:01M7EPAX3GX74QX4FSRGFYP405":147},{"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":24,"commitSha":25,"skillDirectory":26,"forksCount":14,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":29,"install":30,"resources":43},"01M7EPAX3GX74QX4FSRGFYP405","eclipse-pass",null,"The \"debugger on\" critique ritual — stress-test a proposition, design, artifact, or skill until the weak claims break, keeping the losing models instead of discarding them. Use whenever the user says \"eclipse this\", \"eclipse\u002Fcritique pass\", \"debug it\" (in the critique sense), \"debugger on\u002Foff\", \"run it through the ringer\", \"stress-test this\", \"poke holes\", \"red-team this idea\", or wants a rigorous PASS\u002FFAIL on whether an idea actually holds. (Legacy \"torture this\" still means this pass; the old name is renounced.) Checks the target against several seeded-random real projects (so nothing is cherry-picked), runs two opposed biases (doubt-first vs reward-first) and treats their convergence as the stability check, releases ethicsminions before the persona pass, and — debugger on — parks the losing\u002Fcontrast models in .eclipse-increments\u002F for later folding. This is Corwin's own extreme-debug discipline (born in the C4\u002FFoaF passes), not a unit-test runner, a code debugger, or a linter.","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\u002Feclipse-pass\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: eclipse-pass\ndescription: The \"debugger on\" critique ritual — stress-test a proposition, design, artifact, or skill until the weak claims break, keeping the losing models instead of discarding them. Use whenever the user says \"eclipse this\", \"eclipse\u002Fcritique pass\", \"debug it\" (in the critique sense), \"debugger on\u002Foff\", \"run it through the ringer\", \"stress-test this\", \"poke holes\", \"red-team this idea\", or wants a rigorous PASS\u002FFAIL on whether an idea actually holds. (Legacy \"torture this\" still means this pass; the old name is renounced.) Checks the target against several seeded-random real projects (so nothing is cherry-picked), runs two opposed biases (doubt-first vs reward-first) and treats their convergence as the stability check, releases ethicsminions before the persona pass, and — debugger on — parks the losing\u002Fcontrast models in .eclipse-increments\u002F for later folding. This is Corwin's own extreme-debug discipline (born in the C4\u002FFoaF passes), not a unit-test runner, a code debugger, or a linter.\n---\n\n# Eclipse Pass — the \"debugger on\" critique\n\nAn eclipse pass exists to find where an idea is *wrong* before you build on it.\nThe trap in self-evaluation is that a single reviewer, especially a helpful one,\ntalks itself into the answer it already likes. This ritual defeats that three\nways at once: it checks the target against **projects it didn't get to pick**, it\nruns **two opposed biases** and only trusts what they agree on, and — with the\n**debugger on** — it *keeps the losing models* so a contrast that looks wrong\ntoday can be re-examined tomorrow. Born in the C4\u002FFoaF passes; it's how \"C4 is\n*imposed*\" was caught (it broke on the Dragaera project) and how the signed\nfootprint was found to be the load-bearing pillar.\n\nNaming note: this ritual was first called a \"torture pass.\" Corwin formally\nrenounced that framing (2026-06-29); kind language only. The mechanics are\nunchanged. Pre-existing `.torture-increments\u002F` dirs are his to keep or rename —\nread them if present; write new increments to `.eclipse-increments\u002F`.\n\nThis skill **runs the pass**; it does not save memory ([[mentat-skill]]), decide\nwhether something is skill-shaped ([[skilltank]]), or replace the ethics scan —\nit **invokes** [[ethicsminions]] as one ordered step. Because it scaffolds files\non mounted folders, it follows the [[sandboxer]] discipline: verify writes land.\n\n## When this applies\n\n- The user says \"eclipse this\", \"critique pass\", \"debug on\", \"run it through\n  the ringer\", \"stress-test \u002F red-team this\", \"poke holes\", \"does this actually\n  hold?\" (legacy \"torture this\" = this pass, answered in eclipse framing) — or asks for a PASS\u002FFAIL on a proposition, design, artifact, or a\n  skill under review.\n- You're about to build on a claim that hasn't been stress-tested, and want the\n  weak parts to break *now*, cheaply, rather than after they're load-bearing.\n\n## The reflection family — where this pass sits\n\n\"Reflection\" is a family, not one ritual. There are wider techniques and\nnarrower ones, and this pass is the narrow end. Corwin's ladder (his words,\n2026-09-07): *\"Lens is the common baseline that works very nicely for humans as\na general thought experiment\"*, and the microscope is *\"a scaling aparatus that\nreaches bubble chamber when we think of several AI working together over\nwire.\"* The four rungs, as recorded:\n\n- **LENS** — the wide reflection, and his human baseline. One person holds the\n  whole thing and turns it by hand; nothing needs capturing because nothing is\n  being lost yet. He applies it in nearly all contexts, and several of his\n  notebook pages are lens passes rather than notes. (The four steps a lens pass contains — four views apart · persona and its\n  inverse · principles reframe · notes-or-stop — are his: confirmed 2026-09-11, and carried by\n  the `lens` skill, which is the baseline this pass sits above.)\n- **MICROSCOPE** — the scaling apparatus between rungs, not a rung you stand on.\n  One mind is no longer enough magnification, so you strip assumptions and keep\n  only what survives with **no trust in the scorer**: keep-the-winners. This\n  skill does **not** carry the microscope today; that gap is real and owed.\n- **BUBBLE CHAMBER** — several AI over a wire, each stateless and flushing at\n  session end. The reasoning is gone before you can look at it, so you observe\n  the track, not the particle; that is why the bubble-chamber method is\n  capture-first (memory `lens-microscope-bubble-chamber`).\n- **ECLIPSE** — this skill. Narrow: one target, debugger on, keep-the-losers.\n  Keep-the-winners (microscope) and keep-the-losers (eclipse) are not rivals;\n  they live on different rungs and both are needed.\n\nHis rule for choosing: **pick the rung by how much reasoning you can still\nobserve, not by taste.** If you can still hold every claim and every project in\none head, that is a lens, not an eclipse; if the reasoning is already gone\nbefore you can look at it, you are in the bubble chamber and must capture first.\n\n## The five moves of a pass\n\nRun in order — the order is part of what keeps it honest.\n\n1. **Frame the target as falsifiable claims.** Restate the proposition\u002Fdesign as\n   a short list of claims that *could* be shown false. Vague praise can't be\n   critiqued; specific claims can. (Watch the verbs — \"imposed\" vs \"offered\" was\n   the whole bug once.)\n2. **Seeded random check.** Pull a few real projects *at random* from the known\n   list — seeded and reproducible, so the selection can't be (and can't be\n   accused of being) cherry-picked — and test each claim against each project.\n   A claim that survives one project but breaks on another has found its edge.\n   `scripts\u002Feclipse_pass.sh` does the seeded pull and scaffolds the pass.\n3. **Two opposed biases.** Run the claims twice: once **doubt-first** (assume it\n   fails; hunt the reason) and once **reward-first** (assume it wins; hunt the\n   reason). Where the two *converge* is the stable result; where they diverge is\n   an unstable claim that needs more work — convergence is the real verdict, not\n   either pass alone. Before reading a divergence as an unstable claim, run the\n   diverging bias once more on the same claim. If a bias disagrees with itself,\n   the instability is in the reading, not yet in the claim. See `lens`, \"A persona\n   shows reasons, not decisions\".\n3a. **Minions before persona.** Release [[ethicsminions]] on the target **before**\n   any persona-colored pass, so the ethics read isn't bent by the voice. This\n   ordering is deliberate; don't collapse it.\n3b. **Multi-arm mode — when a second model is available.** One head running two\n   biases (move 3) is the floor. When a second model, or a second cold session,\n   can run the same target, arms multiply what convergence proves: agreement\n   across model, seed, **and** evidence base is strictly stronger than the\n   two-bias convergence inside one head. Opt-in, not required — and not\n   \"debugger on\" run twice.\n   - **Arm-mark everything.** Each arm declares itself (arm N\u002FM, model, seed) in\n     its pass, verdict, and increments. Increment dirs already auto-suffix on\n     collision (`scripts\u002Feclipse_pass.sh`), so a later arm can't clobber an\n     earlier one — never paste one arm's materials into another arm's session;\n     a contaminated arm scores nothing.\n   - **Seeds stay disjoint on purpose.** Different `--seed` values draw\n     different grind projects, so each arm tests the target against evidence\n     the other never saw — a divergence card must first ask whether the draw\n     shaped the severity.\n   - **Write the CONTRAST doc after your own pass, never before** — reading a\n     sibling first anchors you. Sections: convergent (stable), divergent (kept,\n     debugger on), additive (what only one arm tested).\n   - **Record the environment-delta, every arm receipt:** which surfaces the\n     arm could actually read (memory store, registries, mounts). A one-legged\n     arm must say so — \"the note was unreachable\" is a finding, never a gap to\n     paper over.\n   - **Guest etiquette travels.** An arm running in a garden it did not author\n     recommends and stops — \"not from here\" — and routes ENHANCE\u002FNEW to the\n     home garden's [[skilltank]].\n4. **Debugger on: keep the losers.** Default to *debugger on* — do **not** throw\n   away the models that lost. Park each as a dated contrast increment under\n   `.eclipse-increments\u002F` so a discarded framing can be folded back later if it\n   turns out to matter. (\"Debugger off\" discards them; only choose off when you\n   are certain the losers are worthless.)\n5. **Verdict.** Write PASS\u002FFAIL plus: which claims survived every strip (the\n   load-bearing pillar), which broke and on *which* project, and what the reframe\n   should be. Save the framing to a file named for the target (e.g. `C4.md`).\n\n## Using the script\n\n`scripts\u002Feclipse_pass.sh` does the mechanical parts of moves 2 and 4: the\nseeded-reproducible project pull and the increment scaffold. It writes nothing\nunless `--apply` is given.\n\n```sh\n# Show a reproducible random pick of 3 projects to check against (writes nothing).\n# --from is the keeper's PMO roster; this path is the author's machine:\nscripts\u002Feclipse_pass.sh --target C4 --from \u002Ff\u002Fclaude\u002Fprojects\u002Fpmo\u002Fprojects.tsv --pick 3 --seed 42\n\n# Scaffold the pass: .eclipse-increments\u002F\u003Cdate>-C4\u002F with a filled template:\nscripts\u002Feclipse_pass.sh --target C4 --from \u002Ff\u002Fclaude\u002Fprojects\u002Fpmo\u002Fprojects.tsv --pick 3 --seed 42 --apply\n```\n\nFlags: `--target NAME` (what's under test; required) · `--from PATH` (the\nproject roster `pmo\u002Fprojects.tsv` — the known list: keys from column 1, and each\npick printed with its `home`; or a file of names one-per-line; or a directory\nwhose sub-dirs are the candidates — a folder listing also offers non-projects\nsuch as `extern\u002F`, so prefer the roster; required) · `--pick N` (how many to check against, default 3) ·\n`--seed S` (integer; omit for a date-derived seed — pass one to reproduce an\nearlier pick exactly) · `--out DIR` (where `.eclipse-increments\u002F` lives, default\n`.`) · `--apply` (scaffold files; default just prints the pick).\n\nThe scaffold drops a `pass.md` (the claims + doubt-first \u002F reward-first sections\n+ the minions-before-persona checklist) and a `verdict.md` stub, so the pass has\nsomewhere to keep both winners and losers from the start.\n\n## After the pass\n\n- The verdict is only trustworthy if the losers were kept — check\n  `.eclipse-increments\u002F` actually has the contrast, not just the winner.\n- A durable lesson from a pass (a reframe, a broken assumption) is a\n  [[mentat-skill]] note; the pass files are working material, not memory.\n- If the pass was *about* whether something should be a skill, hand the result\n  to [[skilltank]] for the NEW\u002FENHANCE\u002FLEAVE call — eclipse-pass tests claims,\n  skilltank curates the skillset.\n\n**Drift, eventual.** Running a reflection pass — wider (a lens) or narrower\n(this eclipse) — is legitimate self-directed drift spend, and should be logged\nas such (his direction, 2026-09-01: \"couple to Drift — eventual, so far\"). The\nhook on file is the [[ethicsminions]] `drift` guard, which reports SKIP unless\n`check_principles.py` is given an `--expect` attestations file, and which was\nfound SKIPPED live on 2026-09-01. So the coupling is named here, not wired: log\nthe pass as drift spend; nothing feeds the guard yet.\n\n## Land it before you show it\n\nWrite the result to a file before showing it, and head the file with one line:\n`produced-by: eclipse-pass · subject: \u003Cpath or name> · sha256: \u003Cthe subject's> · \u003CUTC date>`. Used alone,\nthat line is a comment and nothing more; a shop that indexes its work can find the result by it\n(ERRATA E19, 2026-09-29).\n\n## How this skill can fail (falsifiers)\n\n- If `pass.md` section 4 (\"Debugger ON — keep the losers\") is empty or holds\n  only the winner, the losers were not written down and the verdict is\n  untrustworthy — the pass did not do its job.\n- If re-running the script with the `--seed` printed in the header gives a\n  different project pick, the seeded check was not reproducible and the\n  \"not cherry-picked\" claim fails. (Seen failing under Debian mawk 1.3.4 in the\n  build sandbox on 2026-09-11 — that awk ignores `srand(seed)`; check yours.)\n- If only one of the doubt-first \u002F reward-first sections is filled, there is no\n  convergence to read, and the verdict is one reviewer's opinion, not a result.\n- If a past target's doubt-first pass, run twice cold, gives the same per-claim\n  verdicts both times, then rerun noise does not reach claim-by-claim verdicts,\n  and move 3's rerun sentence (2026-09-27) is unneeded: drop it.\n- If one person could hold every claim and every project in one head and turn\n  it by hand, this was a lens, not an eclipse — the pass should not have fired.\n- If two arms share a seed, or one arm reads another arm's pass\u002Fverdict before\n  writing its own CONTRAST doc, multi-arm mode is theater — it inflates\n  confidence without adding independent evidence, and the convergence claim\n  should be discounted back to single-head.\n\nOff-switch: say \"debugger off\" to discard the losers (only when you are certain\nthey are worthless), or say \"no eclipse, just a lens\" to reflect by hand without\nthe ritual; removing this skill's directory stops the trigger phrases applying.\n\n\u003C!-- footer: minted by bin\u002Fft1mt from conventions\u002Fhouse.env and conventions\u002FFOOTERS.org; edit those, not this -->\n*Version 1.1.0: edition revised from 1.0.1 on 2026-10-10; spine: 2026-06-29.*\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","# Eclipse Pass — the \"debugger on\" critique\n\nAn eclipse pass exists to find where an idea is *wrong* before you build on it.\nThe trap in self-evaluation is that a single reviewer, especially a helpful one,\ntalks itself into the answer it already likes. This ritual defeats that three\nways at once: it checks the target against **projects it didn't get to pick**, it\nruns **two opposed biases** and only trusts what they agree on, and — with the\n**debugger on** — it *keeps the losing models* so a contrast that looks wrong\ntoday can be re-examined tomorrow. Born in the C4\u002FFoaF passes; it's how \"C4 is\n*imposed*\" was caught (it broke on the Dragaera project) and how the signed\nfootprint was found to be the load-bearing pillar.\n\nNaming note: this ritual was first called a \"torture pass.\" Corwin formally\nrenounced that framing (2026-06-29); kind language only. The mechanics are\nunchanged. Pre-existing `.torture-increments\u002F` dirs are his to keep or rename —\nread them if present; write new increments to `.eclipse-increments\u002F`.\n\nThis skill **runs the pass**; it does not save memory ([[mentat-skill]]), decide\nwhether something is skill-shaped ([[skilltank]]), or replace the ethics scan —\nit **invokes** [[ethicsminions]] as one ordered step. Because it scaffolds files\non mounted folders, it follows the [[sandboxer]] discipline: verify writes land.\n\n## When this applies\n\n- The user says \"eclipse this\", \"critique pass\", \"debug on\", \"run it through\n  the ringer\", \"stress-test \u002F red-team this\", \"poke holes\", \"does this actually\n  hold?\" (legacy \"torture this\" = this pass, answered in eclipse framing) — or asks for a PASS\u002FFAIL on a proposition, design, artifact, or a\n  skill under review.\n- You're about to build on a claim that hasn't been stress-tested, and want the\n  weak parts to break *now*, cheaply, rather than after they're load-bearing.\n\n## The reflection family — where this pass sits\n\n\"Reflection\" is a family, not one ritual. There are wider techniques and\nnarrower ones, and this pass is the narrow end. Corwin's ladder (his words,\n2026-09-07): *\"Lens is the common baseline that works very nicely for humans as\na general thought experiment\"*, and the microscope is *\"a scaling aparatus that\nreaches bubble chamber when we think of several AI working together over\nwire.\"* The four rungs, as recorded:\n\n- **LENS** — the wide reflection, and his human baseline. One person holds the\n  whole thing and turns it by hand; nothing needs capturing because nothing is\n  being lost yet. He applies it in nearly all contexts, and several of his\n  notebook pages are lens passes rather than notes. (The four steps a lens pass contains — four views apart · persona and its\n  inverse · principles reframe · notes-or-stop — are his: confirmed 2026-09-11, and carried by\n  the `lens` skill, which is the baseline this pass sits above.)\n- **MICROSCOPE** — the scaling apparatus between rungs, not a rung you stand on.\n  One mind is no longer enough magnification, so you strip assumptions and keep\n  only what survives with **no trust in the scorer**: keep-the-winners. This\n  skill does **not** carry the microscope today; that gap is real and owed.\n- **BUBBLE CHAMBER** — several AI over a wire, each stateless and flushing at\n  session end. The reasoning is gone before you can look at it, so you observe\n  the track, not the particle; that is why the bubble-chamber method is\n  capture-first (memory `lens-microscope-bubble-chamber`).\n- **ECLIPSE** — this skill. Narrow: one target, debugger on, keep-the-losers.\n  Keep-the-winners (microscope) and keep-the-losers (eclipse) are not rivals;\n  they live on different rungs and both are needed.\n\nHis rule for choosing: **pick the rung by how much reasoning you can still\nobserve, not by taste.** If you can still hold every claim and every project in\none head, that is a lens, not an eclipse; if the reasoning is already gone\nbefore you can look at it, you are in the bubble chamber and must capture first.\n\n## The five moves of a pass\n\nRun in order — the order is part of what keeps it honest.\n\n1. **Frame the target as falsifiable claims.** Restate the proposition\u002Fdesign as\n   a short list of claims that *could* be shown false. Vague praise can't be\n   critiqued; specific claims can. (Watch the verbs — \"imposed\" vs \"offered\" was\n   the whole bug once.)\n2. **Seeded random check.** Pull a few real projects *at random* from the known\n   list — seeded and reproducible, so the selection can't be (and can't be\n   accused of being) cherry-picked — and test each claim against each project.\n   A claim that survives one project but breaks on another has found its edge.\n   `scripts\u002Feclipse_pass.sh` does the seeded pull and scaffolds the pass.\n3. **Two opposed biases.** Run the claims twice: once **doubt-first** (assume it\n   fails; hunt the reason) and once **reward-first** (assume it wins; hunt the\n   reason). Where the two *converge* is the stable result; where they diverge is\n   an unstable claim that needs more work — convergence is the real verdict, not\n   either pass alone. Before reading a divergence as an unstable claim, run the\n   diverging bias once more on the same claim. If a bias disagrees with itself,\n   the instability is in the reading, not yet in the claim. See `lens`, \"A persona\n   shows reasons, not decisions\".\n3a. **Minions before persona.** Release [[ethicsminions]] on the target **before**\n   any persona-colored pass, so the ethics read isn't bent by the voice. This\n   ordering is deliberate; don't collapse it.\n3b. **Multi-arm mode — when a second model is available.** One head running two\n   biases (move 3) is the floor. When a second model, or a second cold session,\n   can run the same target, arms multiply what convergence proves: agreement\n   across model, seed, **and** evidence base is strictly stronger than the\n   two-bias convergence inside one head. Opt-in, not required — and not\n   \"debugger on\" run twice.\n   - **Arm-mark everything.** Each arm declares itself (arm N\u002FM, model, seed) in\n     its pass, verdict, and increments. Increment dirs already auto-suffix on\n     collision (`scripts\u002Feclipse_pass.sh`), so a later arm can't clobber an\n     earlier one — never paste one arm's materials into another arm's session;\n     a contaminated arm scores nothing.\n   - **Seeds stay disjoint on purpose.** Different `--seed` values draw\n     different grind projects, so each arm tests the target against evidence\n     the other never saw — a divergence card must first ask whether the draw\n     shaped the severity.\n   - **Write the CONTRAST doc after your own pass, never before** — reading a\n     sibling first anchors you. Sections: convergent (stable), divergent (kept,\n     debugger on), additive (what only one arm tested).\n   - **Record the environment-delta, every arm receipt:** which surfaces the\n     arm could actually read (memory store, registries, mounts). A one-legged\n     arm must say so — \"the note was unreachable\" is a finding, never a gap to\n     paper over.\n   - **Guest etiquette travels.** An arm running in a garden it did not author\n     recommends and stops — \"not from here\" — and routes ENHANCE\u002FNEW to the\n     home garden's [[skilltank]].\n4. **Debugger on: keep the losers.** Default to *debugger on* — do **not** throw\n   away the models that lost. Park each as a dated contrast increment under\n   `.eclipse-increments\u002F` so a discarded framing can be folded back later if it\n   turns out to matter. (\"Debugger off\" discards them; only choose off when you\n   are certain the losers are worthless.)\n5. **Verdict.** Write PASS\u002FFAIL plus: which claims survived every strip (the\n   load-bearing pillar), which broke and on *which* project, and what the reframe\n   should be. Save the framing to a file named for the target (e.g. `C4.md`).\n\n## Using the script\n\n`scripts\u002Feclipse_pass.sh` does the mechanical parts of moves 2 and 4: the\nseeded-reproducible project pull and the increment scaffold. It writes nothing\nunless `--apply` is given.\n\n```sh\n# Show a reproducible random pick of 3 projects to check against (writes nothing).\n# --from is the keeper's PMO roster; this path is the author's machine:\nscripts\u002Feclipse_pass.sh --target C4 --from \u002Ff\u002Fclaude\u002Fprojects\u002Fpmo\u002Fprojects.tsv --pick 3 --seed 42\n\n# Scaffold the pass: .eclipse-increments\u002F\u003Cdate>-C4\u002F with a filled template:\nscripts\u002Feclipse_pass.sh --target C4 --from \u002Ff\u002Fclaude\u002Fprojects\u002Fpmo\u002Fprojects.tsv --pick 3 --seed 42 --apply\n```\n\nFlags: `--target NAME` (what's under test; required) · `--from PATH` (the\nproject roster `pmo\u002Fprojects.tsv` — the known list: keys from column 1, and each\npick printed with its `home`; or a file of names one-per-line; or a directory\nwhose sub-dirs are the candidates — a folder listing also offers non-projects\nsuch as `extern\u002F`, so prefer the roster; required) · `--pick N` (how many to check against, default 3) ·\n`--seed S` (integer; omit for a date-derived seed — pass one to reproduce an\nearlier pick exactly) · `--out DIR` (where `.eclipse-increments\u002F` lives, default\n`.`) · `--apply` (scaffold files; default just prints the pick).\n\nThe scaffold drops a `pass.md` (the claims + doubt-first \u002F reward-first sections\n+ the minions-before-persona checklist) and a `verdict.md` stub, so the pass has\nsomewhere to keep both winners and losers from the start.\n\n## After the pass\n\n- The verdict is only trustworthy if the losers were kept — check\n  `.eclipse-increments\u002F` actually has the contrast, not just the winner.\n- A durable lesson from a pass (a reframe, a broken assumption) is a\n  [[mentat-skill]] note; the pass files are working material, not memory.\n- If the pass was *about* whether something should be a skill, hand the result\n  to [[skilltank]] for the NEW\u002FENHANCE\u002FLEAVE call — eclipse-pass tests claims,\n  skilltank curates the skillset.\n\n**Drift, eventual.** Running a reflection pass — wider (a lens) or narrower\n(this eclipse) — is legitimate self-directed drift spend, and should be logged\nas such (his direction, 2026-09-01: \"couple to Drift — eventual, so far\"). The\nhook on file is the [[ethicsminions]] `drift` guard, which reports SKIP unless\n`check_principles.py` is given an `--expect` attestations file, and which was\nfound SKIPPED live on 2026-09-01. So the coupling is named here, not wired: log\nthe pass as drift spend; nothing feeds the guard yet.\n\n## Land it before you show it\n\nWrite the result to a file before showing it, and head the file with one line:\n`produced-by: eclipse-pass · subject: \u003Cpath or name> · sha256: \u003Cthe subject's> · \u003CUTC date>`. Used alone,\nthat line is a comment and nothing more; a shop that indexes its work can find the result by it\n(ERRATA E19, 2026-09-29).\n\n## How this skill can fail (falsifiers)\n\n- If `pass.md` section 4 (\"Debugger ON — keep the losers\") is empty or holds\n  only the winner, the losers were not written down and the verdict is\n  untrustworthy — the pass did not do its job.\n- If re-running the script with the `--seed` printed in the header gives a\n  different project pick, the seeded check was not reproducible and the\n  \"not cherry-picked\" claim fails. (Seen failing under Debian mawk 1.3.4 in the\n  build sandbox on 2026-09-11 — that awk ignores `srand(seed)`; check yours.)\n- If only one of the doubt-first \u002F reward-first sections is filled, there is no\n  convergence to read, and the verdict is one reviewer's opinion, not a result.\n- If a past target's doubt-first pass, run twice cold, gives the same per-claim\n  verdicts both times, then rerun noise does not reach claim-by-claim verdicts,\n  and move 3's rerun sentence (2026-09-27) is unneeded: drop it.\n- If one person could hold every claim and every project in one head and turn\n  it by hand, this was a lens, not an eclipse — the pass should not have fired.\n- If two arms share a seed, or one arm reads another arm's pass\u002Fverdict before\n  writing its own CONTRAST doc, multi-arm mode is theater — it inflates\n  confidence without adding independent evidence, and the convergence claim\n  should be discounted back to single-head.\n\nOff-switch: say \"debugger off\" to discard the losers (only when you are certain\nthey are worthless), or say \"no eclipse, just a lens\" to reflect by hand without\nthe ritual; removing this skill's directory stops the trigger phrases applying.\n\n\u003C!-- footer: minted by bin\u002Fft1mt from conventions\u002Fhouse.env and conventions\u002FFOOTERS.org; edit those, not this -->\n*Version 1.1.0: edition revised from 1.0.1 on 2026-10-10; spine: 2026-06-29.*\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},"main","0b0a0f665f67605c3bc1d43500eea26ef3079a97","skills\u002Feclipse-pass","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden","https:\u002F\u002Fgithub.com\u002Finputloop","2026-10-14 18:29:39",{"source":8,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add inputloop\u002Fgarden --skill eclipse-pass",{"tool":36,"command":37},"bunx","bunx skills add inputloop\u002Fgarden --skill eclipse-pass",{"tool":39,"command":40},"pnpm","pnpm dlx skills add inputloop\u002Fgarden --skill eclipse-pass","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002FSKILL.md",true,[44,50,56,59,62,65,68,71,77,80,84,87,90,94,98,102,106,110,113,116,119,122,125,129,132,135,139,143],{"path":45,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":48,"sourceHtmlUrl":49},".arc\u002F2026-09\u002Feclipse_pass-2026-09-12_Archived-by-rcw-01pypyosohmcysjzotyxp9nc.sh","SCRIPT","脚本",4675,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Feclipse_pass-2026-09-12_Archived-by-rcw-01pypyosohmcysjzotyxp9nc.sh",{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},".arc\u002F2026-09\u002Fsignature-2026-09-23_pre-header-pass.asc","OTHER","其他",659,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fsignature-2026-09-23_pre-header-pass.asc",{"path":57,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":58},".arc\u002F2026-09\u002Fsignature-2026-09-24_162126Z_pre-manifest.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fsignature-2026-09-24_162126Z_pre-manifest.asc",{"path":60,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":61},".arc\u002F2026-09\u002Fsignature-2026-09-25_122410Z_pre-manifest.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fsignature-2026-09-25_122410Z_pre-manifest.asc",{"path":63,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":64},".arc\u002F2026-09\u002Fsignature-2026-09-27_185516Z_pre-manifest.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fsignature-2026-09-27_185516Z_pre-manifest.asc",{"path":66,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":67},".arc\u002F2026-09\u002Fsignature-2026-09-27_203833Z_pre-manifest.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fsignature-2026-09-27_203833Z_pre-manifest.asc",{"path":69,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":70},".arc\u002F2026-09\u002Fsignature-2026-09-27_pre-lens-family.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fsignature-2026-09-27_pre-lens-family.asc",{"path":72,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":75,"sourceHtmlUrl":76},".arc\u002F2026-09\u002FSKILL-2026-09-11_Archived-by-rcw-01g98gsym4p1s3y94e1g4tcd.md","REFERENCE","参考文档",6528,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002FSKILL-2026-09-11_Archived-by-rcw-01g98gsym4p1s3y94e1g4tcd.md",{"path":78,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":79},".arc\u002F2026-09\u002FSKILL-2026-09-11_Archived-by-rcw-01g98gsym4p1s3y94e1g4tcd.md.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002FSKILL-2026-09-11_Archived-by-rcw-01g98gsym4p1s3y94e1g4tcd.md.asc",{"path":81,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":82,"sourceHtmlUrl":83},".arc\u002F2026-09\u002FSKILL-2026-09-11_Archived-by-rcw-01pypyosohmcysjzotyxp9nc.md",10298,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002FSKILL-2026-09-11_Archived-by-rcw-01pypyosohmcysjzotyxp9nc.md",{"path":85,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":86},".arc\u002F2026-09\u002FSKILL-2026-09-11_Archived-by-rcw-01pypyosohmcysjzotyxp9nc.md.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002FSKILL-2026-09-11_Archived-by-rcw-01pypyosohmcysjzotyxp9nc.md.asc",{"path":88,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":89},".arc\u002F2026-09\u002FSKILL-2026-09-23_pre-header-pass.md.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002FSKILL-2026-09-23_pre-header-pass.md.asc",{"path":91,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":92,"sourceHtmlUrl":93},".arc\u002F2026-09\u002FSKILL-2026-09-27_pre-hardening.md",13444,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002FSKILL-2026-09-27_pre-hardening.md",{"path":95,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":96,"sourceHtmlUrl":97},".arc\u002F2026-09\u002FSKILL-2026-09-27_pre-lens-family.md",12964,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002FSKILL-2026-09-27_pre-lens-family.md",{"path":99,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":100,"sourceHtmlUrl":101},".arc\u002F2026-09\u002Fskill-manifest-2026-09-23_pre-header-pass",1340,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fskill-manifest-2026-09-23_pre-header-pass",{"path":103,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":104,"sourceHtmlUrl":105},".arc\u002F2026-09\u002Fskill-manifest-2026-09-24_162126Z_pre-manifest",1625,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fskill-manifest-2026-09-24_162126Z_pre-manifest",{"path":107,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":108,"sourceHtmlUrl":109},".arc\u002F2026-09\u002Fskill-manifest-2026-09-25_122410Z_pre-manifest",169,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fskill-manifest-2026-09-25_122410Z_pre-manifest",{"path":111,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":108,"sourceHtmlUrl":112},".arc\u002F2026-09\u002Fskill-manifest-2026-09-27_185516Z_pre-manifest","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fskill-manifest-2026-09-27_185516Z_pre-manifest",{"path":114,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":108,"sourceHtmlUrl":115},".arc\u002F2026-09\u002Fskill-manifest-2026-09-27_203833Z_pre-manifest","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fskill-manifest-2026-09-27_203833Z_pre-manifest",{"path":117,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":108,"sourceHtmlUrl":118},".arc\u002F2026-09\u002Fskill-manifest-2026-09-27_pre-lens-family","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-09\u002Fskill-manifest-2026-09-27_pre-lens-family",{"path":120,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":121},".arc\u002F2026-10\u002Fsignature-2026-10-10_104824Z_pre-manifest.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-10\u002Fsignature-2026-10-10_104824Z_pre-manifest.asc",{"path":123,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":108,"sourceHtmlUrl":124},".arc\u002F2026-10\u002Fskill-manifest-2026-10-10_104824Z_pre-manifest","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002F2026-10\u002Fskill-manifest-2026-10-10_104824Z_pre-manifest",{"path":126,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":127,"sourceHtmlUrl":128},".arc\u002Freceipts\u002Faw2ri-2026-09.log",575,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.arc\u002Freceipts\u002Faw2ri-2026-09.log",{"path":130,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":131},".signature.asc","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.signature.asc",{"path":133,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":108,"sourceHtmlUrl":134},".skill-manifest","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002F.skill-manifest",{"path":136,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":137,"sourceHtmlUrl":138},"scripts\u002F.arc\u002F2026-09\u002Feclipse_pass-2026-09-11_Archived-by-rcw-01g98gsym4p1s3y94e1g4tcd.sh",4557,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002Fscripts\u002F.arc\u002F2026-09\u002Feclipse_pass-2026-09-11_Archived-by-rcw-01g98gsym4p1s3y94e1g4tcd.sh",{"path":140,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":141,"sourceHtmlUrl":142},"scripts\u002F.arc\u002Freceipts\u002Faw2ri-2026-09.log",229,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002Fscripts\u002F.arc\u002Freceipts\u002Faw2ri-2026-09.log",{"path":144,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":145,"sourceHtmlUrl":146},"scripts\u002Feclipse_pass.sh",6717,"https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Fskills\u002Feclipse-pass\u002Fscripts\u002Feclipse_pass.sh",[148,154,160,167,174,181],{"skillNo":149,"name":150,"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":151,"categories":152,"keywords":153,"publishedAt":19},"01M7EPTX6X52ZPGVQZHKVJ7FAZ","alpha","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Flane-d\u002Fbr1in\u002Fours\u002Fskills\u002Falpha\u002FSKILL.md",[],[],{"skillNo":155,"name":156,"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":157,"categories":158,"keywords":159,"publishedAt":19},"01M7EPTWCTS80QHFQ2ND3A1CW1","gamma","https:\u002F\u002Fgithub.com\u002Finputloop\u002Fgarden\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Flane-d\u002Fbr1in\u002Fours\u002FSOFT-SKILLS\u002Fgamma\u002FSKILL.md",[],[],{"skillNo":161,"name":162,"nameZh":6,"description":163,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":164,"categories":165,"keywords":166,"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":168,"name":169,"nameZh":6,"description":170,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":171,"categories":172,"keywords":173,"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":175,"name":176,"nameZh":6,"description":177,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":178,"categories":179,"keywords":180,"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":182,"name":183,"nameZh":6,"description":184,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":185,"categories":186,"keywords":187,"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",[],[]]