[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EVCG97X9VD3DB2XY3ZF1SJ":3,"skill-same:01M7EVCG97X9VD3DB2XY3ZF1SJ":46},{"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":29,"ownerDescription":30,"updatedAt":31,"install":32,"resources":45},"01M7EVCG97X9VD3DB2XY3ZF1SJ","fp-finder-skill",null,"Fixed-point convergence engine for skill files — SKILL.md and their references\u002F*.md. Every accepted revision must be strictly shorter (per-file) AND still APPROVING per critical-review, with a factoring escape valve (new sibling whose baseline is bounded by chars moved from a parent that must itself strictly shrink) and a mutual-approval termination capped at N_max iterations. Designed to be dispatched from coordinate-review when the artifact is a SKILL.md, after generic FP loops on SKILL.md files grew unboundedly without converging. Use when you have one or more skill files and want to iterate them to a converged, brief, approved fixed point.","tdupu\u002Fagent-skills-public","tdupu","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F3440403?v=4","01M7EV32X9J0EFE5T4P8TKQTCQ","Taylor Dupuy","01M7EV32XZ1Z4MDBYAFRCJD3BP",0,"GPL-3.0","https:\u002F\u002Fgithub.com\u002Ftdupu\u002Fagent-skills-public\u002Fblob\u002Fmain\u002Fskills\u002Ffp-finder-skill\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: fp-finder-skill\ndescription: Fixed-point convergence engine for skill files — SKILL.md and their references\u002F*.md. Every accepted revision must be strictly shorter (per-file) AND still APPROVING per critical-review, with a factoring escape valve (new sibling whose baseline is bounded by chars moved from a parent that must itself strictly shrink) and a mutual-approval termination capped at N_max iterations. Designed to be dispatched from coordinate-review when the artifact is a SKILL.md, after generic FP loops on SKILL.md files grew unboundedly without converging. Use when you have one or more skill files and want to iterate them to a converged, brief, approved fixed point.\n---\n\n# fp-finder-skill\n\nA skill-file-specialized fixed-point engine (SKILL.md files and the\n`references\u002F*.md` beside them). Every accepted revision must be **strictly shorter per-file** AND **still APPROVING** per `critical-review`. Termination is forced by per-file char-count decrease (Rule a), a parent-shrink + sibling-bound rule on factoring (Rule b), mutual approval, or an `N_max = 8` cap. `N_max = 8` is a guard against runaway iteration; longer-than-8 runs indicate a structural problem the loop alone won't fix and should escalate.\n\n**Char-count is a complexity proxy; APPROVING is the quality floor.** A revision that is shorter but loses APPROVING does not count, so `rm`-style gaming fails: strip the body, the critic rejects, the loop returns the previous best. Per-file (not total-tuple) is the rule everywhere; factoring may grow the tuple, but each file individually still shrinks.\n\n## You are the driver\n\nYou — the agent invoked as `fp-finder-skill` — execute Steps 1–4 below, spawn the critic and revisor subagents, compute char-counts yourself, and decide the verdict on every gate.\n\n## Inputs\n\n- **artifact** (required): a JSON array of one or more skill-file paths — SKILL.md files and\u002For `references\u002F*.md` files under a skill directory — or a single path string. A file with no YAML frontmatter counts its whole content as body (for Rule (b) budgets and critic input alike).\n- **reviewer_persona** (optional): critic lens. Default: \"Skeptical first-time reader of skill files.\"\n\n## State\n\n- `current[path] = (text, wc_c)` — latest accepted state per path. Initialize from inputs with `wc_c = wc -c \u003Cpath>`.\n- `best[path]` — shortest version of `path` that has ever achieved APPROVING **as part of a tuple-level APPROVING round** (Loop Step 3 is the only promotion point; a per-path APPROVING inside a round whose tuple verdict failed never promotes — this keeps `best` tuple-coherent); absent if none.\n- `iteration N`, starts at 1.\n\nAll char-counts are computed by you via `wc -c \u003Cpath>`. The revisor's self-report is reference only.\n\n`normalize(text)` strips trailing whitespace per line and collapses runs of blank lines to one. Used in the \"delta is not whitespace-only\" check.\n\n## Subagent dispatch\n\nSpawn subagents via the Task tool (`general-purpose` on Claude Code; equivalent dispatch primitive elsewhere). If the host environment lacks a dispatcher, run critic and revisor inline as separate prompts in the same session, one after the other. For each subagent skill (`critical-review`, `revise-artifact`), locate the file once via `ls \u003Crepos-root>\u002Fagent-skills\u002Fskills\u002F\u003Cname>\u002FSKILL.md || ls ~\u002F.claude\u002Fskills\u002F\u003Cname>\u002FSKILL.md`, read it, strip the YAML frontmatter, and inline the body in the subagent prompt with the artifact text and any iteration context.\n\n**Tuple aggregation.** For tuples of length > 1, spawn one critic per path. The tuple verdict is APPROVING iff every per-path verdict is APPROVING; otherwise the union of action items (tagged by path) is forwarded to the revisor.\n\n## Loop\n\n**Step 1 — Critic.** Spawn `critical-review` per path; aggregate.\n\n- APPROVING and `N == 1` (initial input never been promoted): for each path, set `best[path] := current[path]`, then go to Step 4 with `initial_input_approved = true`.\n- APPROVING and `N > 1`: go to Step 4.\n- NEEDS-REVISION: collect action items; go to Step 2.\n\n**Step 2 — Revisor with monotone gate.** Spawn `revise-artifact` with the action items AND this **gate instruction verbatim**:\n\n> Your revision must satisfy one branch per modified or new file:\n>\n> (a) **Shrink-in-place.** For each pre-existing file you modify, strictly fewer chars than that file's input version, and the delta must not be whitespace-only.\n>\n> (b) **Factor.** Move a self-contained chunk out of one pre-existing **parent** file into a NEW **sibling** file. The parent MUST end up strictly shorter than its input version (no \"factoring\" without parent-shrink). Per parent in this iteration, the **sum of sibling sizes** must not exceed the chars removed from that parent (so multiple siblings from one parent share a single budget). Per individual sibling: count only the body (everything after the YAML frontmatter) against the budget; the frontmatter is unavoidable overhead and is free.\n>\n> You may combine (a) and (b) across files. Never pad, never copy-shift, never rename to inflate baselines. Report `chars_removed_from_parent` and `sibling_body_wc_c` for every Rule-(b) sibling so the driver can verify.\n\n**Step 3 — Validate then re-critique.** You re-count every modified or new file with `wc -c`. Reject the revision and terminate (return `best`) if any check fails:\n\n| Check | Applies to | Rule |\n|---|---|---|\n| Strict decrease | every modified pre-existing path | `wc_c(new) \u003C current[path].wc_c` |\n| Delta-not-whitespace | every modified file where `wc_c(new) \u003C wc_c(old)` | `normalize(new) != normalize(old)` |\n| Sibling-budget | every parent that produced sibling(s) this iteration | `sum(wc_c(sibling_body)) \u003C= chars_removed_from_parent` |\n| Parent-shrunk on factor | every iteration with one or more new siblings | every named parent satisfies the Strict-decrease check above |\n\nIf all pass, spawn `critical-review` on the new tuple. If APPROVING: promote each changed\u002Fnew path into `current`; update `best[path]` whenever shorter or absent; clear `initial_input_approved` (the input is no longer the fixed point). Increment `N`. If `N > N_max`, go to Wrap-up. Otherwise return to Step 1. If NEEDS-REVISION: revision **fails the floor** — discard and terminate (return `best`).\n\n**Step 4 — Mutual-approval termination.** The critic has approved. Spawn the revisor exactly once with this prompt template (no variant phrasings):\n\n> The artifact tuple below has been approved by the critic. Either:\n> - Answer **NO** if you cannot identify a strict-shrink-or-factor change that would **materially** improve clarity, correctness, or completeness (your contract is to answer NO when no material improvement exists), OR\n> - Provide a concrete revised tuple as a Step-2 revisor would: state action items you want to address, then output the revised file content(s) and any factoring `chars_removed_from_parent` \u002F `sibling_body_wc_c` numbers. Do not answer YES without a concrete revision payload.\n>\n> [current tuple]\n\nNO → mutual approval; return `best`. Concrete revision → treat as Step 2 output (re-validate in Step 3) and continue.\n\n## Termination\n\nEvery accepted revision strictly decreases at least one file's char-count (Rule (a) decreases the modified file; Rule (b) decreases the parent via the Parent-shrunk-on-factor check). Char-counts are non-negative integers, so per-file budgets sum to finite. Factoring introduces new siblings, but the per-parent sibling-budget bounds the sum of sibling bodies by chars-removed from that parent — so the total system-wide body chars never grow more than they shrank on the same iteration. Frontmatter overhead per sibling is constant; total siblings created across all iterations is bounded by initial total parent body chars divided by the smallest meaningful factor, which is finite. `N_max = 8` is the explicit backstop.\n\n## Returns\n\n- `best[path]` present: shortest APPROVING version of that path.\n- `best[path]` absent (initial input never APPROVING, or a revision lost the floor before any APPROVING pass): return the path's initial text with note `\"no improvement confirmed.\"`\n- `initial_input_approved = true` at termination: prefix `best` with note `\"converged at N=1; initial input was already a fixed point.\"`\n- Factored-out parents: their `best` is the APPROVING version of the now-narrower parent.\n- Siblings created in an iteration whose re-critic then returned NEEDS-REVISION: omit from `best`; return their text-on-disk with note `\"sibling created but iteration discarded; review manually.\"`\n- `N_max` hit: prefix wrap-up with `\"stagnated at N_max iterations; returning best so far.\"`\n\n## Pitfalls\n\n- **Padding gaming.** Step 3's delta-not-whitespace + the critic's bloat detection together defend against whitespace and meaningless-token padding.\n- **Spurious factoring.** A sibling with no consumer is dispatch overhead. The factored chunk must be self-contained, the parent's narrowed `description` must still cover it, and the parent should reference the new sibling.\n- **Revisor lying about chars.** You always re-count via `wc -c`. The revisor cannot evade.\n- **Char-count rewards terseness over examples.** Concrete examples may lose to abstract prose. The APPROVING floor protects against this: load-bearing examples removed → critic blocks → revision loses the floor.\n- **The proxy is not the goal.** Minimize char-count *subject to* APPROVING; never for its own sake.\n","# fp-finder-skill\n\nA skill-file-specialized fixed-point engine (SKILL.md files and the\n`references\u002F*.md` beside them). Every accepted revision must be **strictly shorter per-file** AND **still APPROVING** per `critical-review`. Termination is forced by per-file char-count decrease (Rule a), a parent-shrink + sibling-bound rule on factoring (Rule b), mutual approval, or an `N_max = 8` cap. `N_max = 8` is a guard against runaway iteration; longer-than-8 runs indicate a structural problem the loop alone won't fix and should escalate.\n\n**Char-count is a complexity proxy; APPROVING is the quality floor.** A revision that is shorter but loses APPROVING does not count, so `rm`-style gaming fails: strip the body, the critic rejects, the loop returns the previous best. Per-file (not total-tuple) is the rule everywhere; factoring may grow the tuple, but each file individually still shrinks.\n\n## You are the driver\n\nYou — the agent invoked as `fp-finder-skill` — execute Steps 1–4 below, spawn the critic and revisor subagents, compute char-counts yourself, and decide the verdict on every gate.\n\n## Inputs\n\n- **artifact** (required): a JSON array of one or more skill-file paths — SKILL.md files and\u002For `references\u002F*.md` files under a skill directory — or a single path string. A file with no YAML frontmatter counts its whole content as body (for Rule (b) budgets and critic input alike).\n- **reviewer_persona** (optional): critic lens. Default: \"Skeptical first-time reader of skill files.\"\n\n## State\n\n- `current[path] = (text, wc_c)` — latest accepted state per path. Initialize from inputs with `wc_c = wc -c \u003Cpath>`.\n- `best[path]` — shortest version of `path` that has ever achieved APPROVING **as part of a tuple-level APPROVING round** (Loop Step 3 is the only promotion point; a per-path APPROVING inside a round whose tuple verdict failed never promotes — this keeps `best` tuple-coherent); absent if none.\n- `iteration N`, starts at 1.\n\nAll char-counts are computed by you via `wc -c \u003Cpath>`. The revisor's self-report is reference only.\n\n`normalize(text)` strips trailing whitespace per line and collapses runs of blank lines to one. Used in the \"delta is not whitespace-only\" check.\n\n## Subagent dispatch\n\nSpawn subagents via the Task tool (`general-purpose` on Claude Code; equivalent dispatch primitive elsewhere). If the host environment lacks a dispatcher, run critic and revisor inline as separate prompts in the same session, one after the other. For each subagent skill (`critical-review`, `revise-artifact`), locate the file once via `ls \u003Crepos-root>\u002Fagent-skills\u002Fskills\u002F\u003Cname>\u002FSKILL.md || ls ~\u002F.claude\u002Fskills\u002F\u003Cname>\u002FSKILL.md`, read it, strip the YAML frontmatter, and inline the body in the subagent prompt with the artifact text and any iteration context.\n\n**Tuple aggregation.** For tuples of length > 1, spawn one critic per path. The tuple verdict is APPROVING iff every per-path verdict is APPROVING; otherwise the union of action items (tagged by path) is forwarded to the revisor.\n\n## Loop\n\n**Step 1 — Critic.** Spawn `critical-review` per path; aggregate.\n\n- APPROVING and `N == 1` (initial input never been promoted): for each path, set `best[path] := current[path]`, then go to Step 4 with `initial_input_approved = true`.\n- APPROVING and `N > 1`: go to Step 4.\n- NEEDS-REVISION: collect action items; go to Step 2.\n\n**Step 2 — Revisor with monotone gate.** Spawn `revise-artifact` with the action items AND this **gate instruction verbatim**:\n\n> Your revision must satisfy one branch per modified or new file:\n>\n> (a) **Shrink-in-place.** For each pre-existing file you modify, strictly fewer chars than that file's input version, and the delta must not be whitespace-only.\n>\n> (b) **Factor.** Move a self-contained chunk out of one pre-existing **parent** file into a NEW **sibling** file. The parent MUST end up strictly shorter than its input version (no \"factoring\" without parent-shrink). Per parent in this iteration, the **sum of sibling sizes** must not exceed the chars removed from that parent (so multiple siblings from one parent share a single budget). Per individual sibling: count only the body (everything after the YAML frontmatter) against the budget; the frontmatter is unavoidable overhead and is free.\n>\n> You may combine (a) and (b) across files. Never pad, never copy-shift, never rename to inflate baselines. Report `chars_removed_from_parent` and `sibling_body_wc_c` for every Rule-(b) sibling so the driver can verify.\n\n**Step 3 — Validate then re-critique.** You re-count every modified or new file with `wc -c`. Reject the revision and terminate (return `best`) if any check fails:\n\n| Check | Applies to | Rule |\n|---|---|---|\n| Strict decrease | every modified pre-existing path | `wc_c(new) \u003C current[path].wc_c` |\n| Delta-not-whitespace | every modified file where `wc_c(new) \u003C wc_c(old)` | `normalize(new) != normalize(old)` |\n| Sibling-budget | every parent that produced sibling(s) this iteration | `sum(wc_c(sibling_body)) \u003C= chars_removed_from_parent` |\n| Parent-shrunk on factor | every iteration with one or more new siblings | every named parent satisfies the Strict-decrease check above |\n\nIf all pass, spawn `critical-review` on the new tuple. If APPROVING: promote each changed\u002Fnew path into `current`; update `best[path]` whenever shorter or absent; clear `initial_input_approved` (the input is no longer the fixed point). Increment `N`. If `N > N_max`, go to Wrap-up. Otherwise return to Step 1. If NEEDS-REVISION: revision **fails the floor** — discard and terminate (return `best`).\n\n**Step 4 — Mutual-approval termination.** The critic has approved. Spawn the revisor exactly once with this prompt template (no variant phrasings):\n\n> The artifact tuple below has been approved by the critic. Either:\n> - Answer **NO** if you cannot identify a strict-shrink-or-factor change that would **materially** improve clarity, correctness, or completeness (your contract is to answer NO when no material improvement exists), OR\n> - Provide a concrete revised tuple as a Step-2 revisor would: state action items you want to address, then output the revised file content(s) and any factoring `chars_removed_from_parent` \u002F `sibling_body_wc_c` numbers. Do not answer YES without a concrete revision payload.\n>\n> [current tuple]\n\nNO → mutual approval; return `best`. Concrete revision → treat as Step 2 output (re-validate in Step 3) and continue.\n\n## Termination\n\nEvery accepted revision strictly decreases at least one file's char-count (Rule (a) decreases the modified file; Rule (b) decreases the parent via the Parent-shrunk-on-factor check). Char-counts are non-negative integers, so per-file budgets sum to finite. Factoring introduces new siblings, but the per-parent sibling-budget bounds the sum of sibling bodies by chars-removed from that parent — so the total system-wide body chars never grow more than they shrank on the same iteration. Frontmatter overhead per sibling is constant; total siblings created across all iterations is bounded by initial total parent body chars divided by the smallest meaningful factor, which is finite. `N_max = 8` is the explicit backstop.\n\n## Returns\n\n- `best[path]` present: shortest APPROVING version of that path.\n- `best[path]` absent (initial input never APPROVING, or a revision lost the floor before any APPROVING pass): return the path's initial text with note `\"no improvement confirmed.\"`\n- `initial_input_approved = true` at termination: prefix `best` with note `\"converged at N=1; initial input was already a fixed point.\"`\n- Factored-out parents: their `best` is the APPROVING version of the now-narrower parent.\n- Siblings created in an iteration whose re-critic then returned NEEDS-REVISION: omit from `best`; return their text-on-disk with note `\"sibling created but iteration discarded; review manually.\"`\n- `N_max` hit: prefix wrap-up with `\"stagnated at N_max iterations; returning best so far.\"`\n\n## Pitfalls\n\n- **Padding gaming.** Step 3's delta-not-whitespace + the critic's bloat detection together defend against whitespace and meaningless-token padding.\n- **Spurious factoring.** A sibling with no consumer is dispatch overhead. The factored chunk must be self-contained, the parent's narrowed `description` must still cover it, and the parent should reference the new sibling.\n- **Revisor lying about chars.** You always re-count via `wc -c`. The revisor cannot evade.\n- **Char-count rewards terseness over examples.** Concrete examples may lose to abstract prose. The APPROVING floor protects against this: load-bearing examples removed → critic blocks → revision loses the floor.\n- **The proxy is not the goal.** Minimize char-count *subject to* APPROVING; never for its own sake.\n",{"name":5,"description":7},"main","8d13efb5e66306aeae6e3b948bc463e7e2db1b65","skills\u002Ffp-finder-skill","https:\u002F\u002Fgithub.com\u002Ftdupu\u002Fagent-skills-public","https:\u002F\u002Fgithub.com\u002Ftdupu","https:\u002F\u002Ftdupu.github.io\u002Findex.html","Associate Professor, Department of Mathematics and Statistics, University of Vermont --- last updated Fall 2019","2026-11-13 09:07:19",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add tdupu\u002Fagent-skills-public --skill fp-finder-skill",{"tool":38,"command":39},"bunx","bunx skills add tdupu\u002Fagent-skills-public --skill fp-finder-skill",{"tool":41,"command":42},"pnpm","pnpm dlx skills add tdupu\u002Fagent-skills-public --skill fp-finder-skill","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Ftdupu\u002Fagent-skills-public\u002Fblob\u002Fmain\u002Fskills\u002Ffp-finder-skill\u002FSKILL.md",true,[],[47,54,61,68,75,82],{"skillNo":48,"name":49,"nameZh":6,"description":50,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":51,"categories":52,"keywords":53,"publishedAt":19},"01M7EW4PT586KKE7VRCAD80Z9K","writing-skills","Use when creating new skills, editing existing skills, or verifying skills work before deployment","https:\u002F\u002Fgithub.com\u002Ftdupu\u002Fagent-skills-public\u002Fblob\u002Fmain\u002Fsuperpowers\u002Fskills\u002Fwriting-skills\u002FSKILL.md",[],[],{"skillNo":55,"name":56,"nameZh":6,"description":57,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":58,"categories":59,"keywords":60,"publishedAt":19},"01M7EW4HV5V74AF3PTV8N4T0PS","writing-plans","Use when you have a spec or requirements for a multi-step task, before touching code","https:\u002F\u002Fgithub.com\u002Ftdupu\u002Fagent-skills-public\u002Fblob\u002Fmain\u002Fsuperpowers\u002Fskills\u002Fwriting-plans\u002FSKILL.md",[],[],{"skillNo":62,"name":63,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":19},"01M7EW4DNKGP5CR1RAJP2J9PHA","verification-before-completion","Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always","https:\u002F\u002Fgithub.com\u002Ftdupu\u002Fagent-skills-public\u002Fblob\u002Fmain\u002Fsuperpowers\u002Fskills\u002Fverification-before-completion\u002FSKILL.md",[],[],{"skillNo":69,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":19},"01M7EW4BA2QP88VBFVXDVJM7XQ","using-superpowers","Use when starting any conversation - establishes how to find and use skills, requiring skill invocation before ANY response including clarifying questions","https:\u002F\u002Fgithub.com\u002Ftdupu\u002Fagent-skills-public\u002Fblob\u002Fmain\u002Fsuperpowers\u002Fskills\u002Fusing-superpowers\u002FSKILL.md",[],[],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":19},"01M7EW47F2JG5FBESDJNP3PKMV","using-git-worktrees","Use when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback","https:\u002F\u002Fgithub.com\u002Ftdupu\u002Fagent-skills-public\u002Fblob\u002Fmain\u002Fsuperpowers\u002Fskills\u002Fusing-git-worktrees\u002FSKILL.md",[],[],{"skillNo":83,"name":84,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":19},"01M7EW44N8556Y556NX9HJT4KP","test-driven-development","Use when implementing any feature or bugfix, before writing implementation code","https:\u002F\u002Fgithub.com\u002Ftdupu\u002Fagent-skills-public\u002Fblob\u002Fmain\u002Fsuperpowers\u002Fskills\u002Ftest-driven-development\u002FSKILL.md",[],[]]