[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77H9BYAXMY7M0AVNYG0S2BC":3,"skill-same:01M77H9BYAXMY7M0AVNYG0S2BC":50},{"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":27,"commitSha":28,"skillDirectory":29,"forksCount":30,"repositoryHtmlUrl":31,"repositoryHomepageUrl":32,"ownerHtmlUrl":33,"ownerBlogUrl":34,"ownerDescription":6,"updatedAt":35,"install":36,"resources":49},"01M77H9BYAXMY7M0AVNYG0S2BC","token-defaults-score",null,"One repeatable pass that keeps fak's OUT-OF-THE-BOX token economy amazing — every stacking token-saving method that can SAFELY default is on by default, honestly noted, and locked against regression. Runs the token-saving-defaults scorecard (`fak token-defaults-scorecard`) over the entrypoint source (cmd\u002Ffak\u002Fguard.go, cmd\u002Ffak\u002Fserve.go, the gateway Default* constants, and the audited servewiringData rows), turns each HARD defect into a required fix — turn a WITNESSED-safe bounded-loss saver on by default with an honest note, write the missing regression lock, document a dark lever's. Use when this named workflow matches the task.","anthony-chaudhary\u002Ffak","anthony-chaudhary","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F18080164?v=4","01M77GSKCAEKV3JS07J40B94DY","Anthony Chaudhary","01M77GSKD26Z7F4FQQD4VA30VX",48,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Ftoken-defaults-score\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: token-defaults-score\ndescription: One repeatable pass that keeps fak's OUT-OF-THE-BOX token economy amazing — every stacking token-saving method that can SAFELY default is on by default, honestly noted, and locked against regression. Runs the token-saving-defaults scorecard (`fak token-defaults-scorecard`) over the entrypoint source (cmd\u002Ffak\u002Fguard.go, cmd\u002Ffak\u002Fserve.go, the gateway Default* constants, and the audited servewiringData rows), turns each HARD defect into a required fix — turn a WITNESSED-safe bounded-loss saver on by default with an honest note, write the missing regression lock, document a dark lever's. Use when this named workflow matches the task.\nmetadata:\n  generated-by: fak project-assets sync\n  canonical: ..\u002F..\u002F..\u002F.claude\u002Fskills\u002Ftoken-defaults-score\u002FSKILL.md\n---\n\n# token-defaults-score — keep the out-of-the-box token economy amazing\n\nThis is an instance of the **`scorecard` doctrine** (`.claude\u002Fskills\u002Fscorecard\u002FSKILL.md`) —\nread that for the shared five laws and the anti-gaming rule. This file is the per-surface\nloop for **fak's out-of-the-box token-saving defaults**: of every stacking token-saving\nmethod fak knows, which are ON by default when a user runs `fak guard -- claude` \u002F `fak\nserve` with no flags — and are the high-value, low-loss ones turned on out of the box, each\nhonestly noted, none able to regress unnoticed?\n\nThe headline metric is **token-defaults-debt** (`corpus.token_defaults_debt`). Drive it to\nzero, then — per the `score-2x` doctrine — HARDEN: produce the witness that promotes the\nnext off lever to a must-default.\n\n## The honesty law that makes this scorecard different\n\nA token saver belongs ON by default the moment it is **demonstrably** safe — the operator's\n\"if it keeps ~the model's view and saves big, default it\" rule. **Demonstrably** is the\nload-bearing word: turning an *unwitnessed* lever on would ship an unproven claim, which\n[`CLAIMS.md`](..\u002F..\u002F..\u002FCLAIMS.md) \u002F [`CLAUDE.md`](..\u002F..\u002F..\u002FCLAUDE.md) forbid. So:\n\n- A **lossless** saver (provider-cache passthrough, tool-floor pruning, vDSO dedup) has no\n  fidelity tradeoff → it MUST be on; off is HARD debt.\n- A **bounded-loss** saver (history compaction, oversized-result elision) is high-value, but\n  it is hard debt to leave off ONLY when a committed **witness** quantifies its\n  savings\u002Ffidelity tradeoff AND no gate remains. Unwitnessed → it is a SOFT \"needs a\n  witness,\" never a forced flip.\n- An **opt-in** saver (the ctxplan view) stays off behind a documented gate; its witness\n  (e.g. 13.3× fewer resident, 100% exact recall) makes it the next default to turn on once\n  the gate (a watched-live session \u002F live-loop wiring) clears.\n\nThe scorecard reads the REAL default from the entrypoint source and cross-checks the witness\nagainst committed docs — you cannot drop debt by editing the roster. **Never turn an\nunwitnessed lever on to score.** Produce the witness first.\n\n## The loop (the shared five steps, this surface)\n\n1. **Run it** — `go run .\u002Fcmd\u002Ffak token-defaults-scorecard` (human work-list);\n   `--json` (machine payload); `--markdown` (the committed snapshot). Read the **per-lever\n   status table** first — it's the \"where each saver stands\" view (default, witness,\n   blocker, noted, locked).\n\n2. **Retire token-defaults-debt worst-first** — by changing reality:\n   - `lossless_stack` \u002F `high_value_defaults` defect → turn the saver ON by default (flip the\n     `Default*` const \u002F the flag literal in both `guard.go` and `serve.go`), add an honest\n     loss note to the flag help, and regenerate the serve-wiring row (`fak serve-wiring --md`)\n     so the audited verdict tracks the flip.\n   - `default_on_locked` defect → add a real assertion to `cmd\u002Ffak\u002Ftoken_defaults_test.go`\n     that pins the default (reds if a peer flips it back).\n   - `dark_lever_gated` defect → document why the lever is dark + a safe value + the witness\n     it needs, in the flag help \u002F const comment.\n   - `entrypoint_parity` defect → make `guard.go` and `serve.go` agree, or regenerate the\n     stale servewiring verdict.\n\n3. **Weigh the SOFT signals, then stop** — `witness_status` is the roadmap, not a fix: an\n   *unwitnessed* high-value lever needs a measurement (build\u002Frun the witness — e.g. a\n   deterministic shed-ratio replay, or the watched-live dogfood the gate names), and a\n   *witnessed-gated* lever needs its gate cleared. Producing a witness is real work; do it\n   deliberately, then the next run promotes that lever to a must-default.\n\n4. **Re-measure + prove** — rerun\n   `go run .\u002Fcmd\u002Ffak token-defaults-scorecard --json` and compare the debt against the\n   baseline you recorded; regenerate the snapshot\n   (`go run .\u002Fcmd\u002Ffak token-defaults-scorecard --markdown > docs\u002Fserving\u002Ftoken-defaults-scorecard.md`);\n   run the locks (`go test .\u002Fcmd\u002Ffak -run TestTokenDefault`).\n\n5. **Commit only the scorecard lane, by explicit path** — run `fak sync reconcile --apply`, then commit via `fak commit --path` (never `git add -A`). End the\n   subject with the `(fak \u003Cleaf>)` trailer (stamp the dominant leaf — `tools` for a\n   scorecard-only pass, `gateway`\u002F`cmd` when the same pass flips a default). Push via `fak sync push` or `[--push]`. e.g.\n   ```bash\n   fak sync reconcile --apply\n   fak commit --path cmd\u002Ffak\u002Ftoken_defaults.go --path cmd\u002Ffak\u002Ftoken_defaults_test.go --path docs\u002Fserving\u002Ftoken-defaults-scorecard.md -F msg [--push]\n   fak sync push\n   ```\n\n## The anti-gaming rule (this surface)\n\nRetire a defect by changing the binary's real default \u002F writing a real lock \u002F producing a\nreal witness — never by editing the roster, loosening the witness regex, or pointing a\nsentinel at a vacuous test. A `default_on_locked` defect is fixed by a test that actually\nasserts the default and reds when it flips, not by a test that merely names the lever. The\none move that turns this into theater is turning a lever on to score before it is witnessed:\nthat ships exactly the unproven claim the whole repo refuses.\n","# token-defaults-score — keep the out-of-the-box token economy amazing\n\nThis is an instance of the **`scorecard` doctrine** (`.claude\u002Fskills\u002Fscorecard\u002FSKILL.md`) —\nread that for the shared five laws and the anti-gaming rule. This file is the per-surface\nloop for **fak's out-of-the-box token-saving defaults**: of every stacking token-saving\nmethod fak knows, which are ON by default when a user runs `fak guard -- claude` \u002F `fak\nserve` with no flags — and are the high-value, low-loss ones turned on out of the box, each\nhonestly noted, none able to regress unnoticed?\n\nThe headline metric is **token-defaults-debt** (`corpus.token_defaults_debt`). Drive it to\nzero, then — per the `score-2x` doctrine — HARDEN: produce the witness that promotes the\nnext off lever to a must-default.\n\n## The honesty law that makes this scorecard different\n\nA token saver belongs ON by default the moment it is **demonstrably** safe — the operator's\n\"if it keeps ~the model's view and saves big, default it\" rule. **Demonstrably** is the\nload-bearing word: turning an *unwitnessed* lever on would ship an unproven claim, which\n[`CLAIMS.md`](..\u002F..\u002F..\u002FCLAIMS.md) \u002F [`CLAUDE.md`](..\u002F..\u002F..\u002FCLAUDE.md) forbid. So:\n\n- A **lossless** saver (provider-cache passthrough, tool-floor pruning, vDSO dedup) has no\n  fidelity tradeoff → it MUST be on; off is HARD debt.\n- A **bounded-loss** saver (history compaction, oversized-result elision) is high-value, but\n  it is hard debt to leave off ONLY when a committed **witness** quantifies its\n  savings\u002Ffidelity tradeoff AND no gate remains. Unwitnessed → it is a SOFT \"needs a\n  witness,\" never a forced flip.\n- An **opt-in** saver (the ctxplan view) stays off behind a documented gate; its witness\n  (e.g. 13.3× fewer resident, 100% exact recall) makes it the next default to turn on once\n  the gate (a watched-live session \u002F live-loop wiring) clears.\n\nThe scorecard reads the REAL default from the entrypoint source and cross-checks the witness\nagainst committed docs — you cannot drop debt by editing the roster. **Never turn an\nunwitnessed lever on to score.** Produce the witness first.\n\n## The loop (the shared five steps, this surface)\n\n1. **Run it** — `go run .\u002Fcmd\u002Ffak token-defaults-scorecard` (human work-list);\n   `--json` (machine payload); `--markdown` (the committed snapshot). Read the **per-lever\n   status table** first — it's the \"where each saver stands\" view (default, witness,\n   blocker, noted, locked).\n\n2. **Retire token-defaults-debt worst-first** — by changing reality:\n   - `lossless_stack` \u002F `high_value_defaults` defect → turn the saver ON by default (flip the\n     `Default*` const \u002F the flag literal in both `guard.go` and `serve.go`), add an honest\n     loss note to the flag help, and regenerate the serve-wiring row (`fak serve-wiring --md`)\n     so the audited verdict tracks the flip.\n   - `default_on_locked` defect → add a real assertion to `cmd\u002Ffak\u002Ftoken_defaults_test.go`\n     that pins the default (reds if a peer flips it back).\n   - `dark_lever_gated` defect → document why the lever is dark + a safe value + the witness\n     it needs, in the flag help \u002F const comment.\n   - `entrypoint_parity` defect → make `guard.go` and `serve.go` agree, or regenerate the\n     stale servewiring verdict.\n\n3. **Weigh the SOFT signals, then stop** — `witness_status` is the roadmap, not a fix: an\n   *unwitnessed* high-value lever needs a measurement (build\u002Frun the witness — e.g. a\n   deterministic shed-ratio replay, or the watched-live dogfood the gate names), and a\n   *witnessed-gated* lever needs its gate cleared. Producing a witness is real work; do it\n   deliberately, then the next run promotes that lever to a must-default.\n\n4. **Re-measure + prove** — rerun\n   `go run .\u002Fcmd\u002Ffak token-defaults-scorecard --json` and compare the debt against the\n   baseline you recorded; regenerate the snapshot\n   (`go run .\u002Fcmd\u002Ffak token-defaults-scorecard --markdown > docs\u002Fserving\u002Ftoken-defaults-scorecard.md`);\n   run the locks (`go test .\u002Fcmd\u002Ffak -run TestTokenDefault`).\n\n5. **Commit only the scorecard lane, by explicit path** — run `fak sync reconcile --apply`, then commit via `fak commit --path` (never `git add -A`). End the\n   subject with the `(fak \u003Cleaf>)` trailer (stamp the dominant leaf — `tools` for a\n   scorecard-only pass, `gateway`\u002F`cmd` when the same pass flips a default). Push via `fak sync push` or `[--push]`. e.g.\n   ```bash\n   fak sync reconcile --apply\n   fak commit --path cmd\u002Ffak\u002Ftoken_defaults.go --path cmd\u002Ffak\u002Ftoken_defaults_test.go --path docs\u002Fserving\u002Ftoken-defaults-scorecard.md -F msg [--push]\n   fak sync push\n   ```\n\n## The anti-gaming rule (this surface)\n\nRetire a defect by changing the binary's real default \u002F writing a real lock \u002F producing a\nreal witness — never by editing the roster, loosening the witness regex, or pointing a\nsentinel at a vacuous test. A `default_on_locked` defect is fixed by a test that actually\nasserts the default and reds when it flips, not by a test that merely names the lever. The\none move that turns this into theater is turning a lever on to score before it is witnessed:\nthat ships exactly the unproven claim the whole repo refuses.\n",{"name":5,"metadata":24,"description":7},{"canonical":25,"generated-by":26},"..\u002F..\u002F..\u002F.claude\u002Fskills\u002Ftoken-defaults-score\u002FSKILL.md","fak project-assets sync","main","f565b7b7db63e79649f3a02c08108132a86c5d05",".agents\u002Fskills\u002Ftoken-defaults-score",16,"https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak","https:\u002F\u002Fanthony-chaudhary.github.io\u002Ffak\u002Fshowcase.html","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary","https:\u002F\u002Flinkedin.com\u002Fin\u002Fanthonychaudhary","2026-11-15 03:06:00",{"source":8,"skillName":5,"commands":37,"prompt":47,"zipAvailable":48},[38,41,44],{"tool":39,"command":40},"npx","npx skills add anthony-chaudhary\u002Ffak --skill token-defaults-score",{"tool":42,"command":43},"bunx","bunx skills add anthony-chaudhary\u002Ffak --skill token-defaults-score",{"tool":45,"command":46},"pnpm","pnpm dlx skills add anthony-chaudhary\u002Ffak --skill token-defaults-score","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Ftoken-defaults-score\u002FSKILL.md",true,[],[51,58,67,74,81,86],{"skillNo":52,"name":53,"nameZh":6,"description":54,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":55,"categories":56,"keywords":57,"publishedAt":19},"01M78V4059KRP2T88YR6Z05MJQ","serve-receipt","On-demand research protocol + deterministic receipt for serving a Hugging Face model on a concrete GPU target — fan out research lanes (model facts\u002Ffiles, engine support matrix, serving recipes, KV\u002Fstate memory math), compute the tiered feasibility verdict with `fak serve-receipt` (never narrated math), and emit a witnessed launch receipt. Use when asked for an \"initial config\", \"serving receipt\", or \"how do I serve \u003Cmodel> on \u003CGPUs>\".","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fserve-receipt\u002FSKILL.md",[],[],{"skillNo":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":62,"licenseFileUrl":63,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M77HRGKQV2CJ2K6ZFN4YQ32K","caveman","Ultra-compressed communication mode. Cuts output tokens 65% (measured) by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says \"caveman mode\", \"talk like caveman\", \"use caveman\", \"less tokens\", \"be brief\", or invokes \u002Fcaveman. Also auto-triggers when token efficiency is requested.","MIT","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fsyspromptmmu\u002Fthird_party\u002Fcaveman\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fsyspromptmmu\u002Fthird_party\u002Fcaveman\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":19},"01M77HRAY3Y3BE7APN5PVJF7NV","quality-report","Score the repo quality and emit a report card.","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplans\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fquality-report\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M77HR9807KHAX13DG9PW2K2Q","ctx-overlay","Compact the context window and prune the memory store when a session grows too long.","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplans\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fctx-overlay\u002FSKILL.md",[],[],{"skillNo":82,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":19},"01M77HR7XQ9X8F2CY9Z4F2ZEZ3","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplanlint\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fquality-report\u002FSKILL.md",[],[],{"skillNo":87,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":88,"categories":89,"keywords":90,"publishedAt":19},"01M77HR5BVKCTR1WNJGRBSDNX7","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplanlint\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fctx-overlay\u002FSKILL.md",[],[]]