All authors
Claude Skills by p2p2sp
github.com/p2p2sp61 skills0 installs51 views
- AdrJudges an intent interview's confirmed decisions against the three ADR criteria (hard to reverse, surprising without context, the result of a real trade-off) and returns a one-paragraph architecture decision record for each decision that passes all three. Invoked by the `intent` skill only - never directly, never by the user, never mid-interview, never from a spec, plan, build or review.Votes: 0GitHub stars: 8
- E2eGenerate and locally verify Playwright tests for the scenarios of one docs/qa handoff file, then commit them for CI.Votes: 0GitHub stars: 8
- ExecutorRuns ONE build, test, lint, type-check or other shell command in a forked context and returns a short verdict - the tool's own summary line, the failures, a log path - instead of the output itself. Use it whenever an outcome must be judged but its output must not enter the caller's context. Two modes, one reply shape: `command:` runs the command verbatim; `log:` + `exit:` + `duration:` runs nothing and reads a log an earlier direct `run.sh` call already wrote. Not for output the caller wants ...Votes: 0GitHub stars: 8
- IntentYou MUST ALWAYS use this skill every time a user wants to do something creative - a new idea, a new feature, build something from scratch, a change to an existing solution.Votes: 0GitHub stars: 8
- Phases ReviewerInvoked only by phases skill.Votes: 0GitHub stars: 8
- PhasesCuts one confirmed intent into ordered phases, saves phases.md plus one intent.md per phase, and resumes a phased run. Use when an intent is too large for a single spec or plan, when the user asks to split the work into phases, stages, milestones or chunks, when they ask which phase comes next, what the phase status is, or to resume, continue or reopen a phased undertaking. Invoked from the intent skill's handoff gate (Phases option) or by the user command `phases` on an existing intent.md or...Votes: 0GitHub stars: 8
- SetupSetup superdev environment.Votes: 0GitHub stars: 8
- SimplebuildBuild orchestrator for an approved SimplePlan (a plan without a Spec line). Use it ONLY when the approved plan's body says to build it with the simplebuild skill - never for a plan that names superbuild, never without an approved plan.Votes: 0GitHub stars: 8
- SimpledebugUse when the user reports a bug, error, crash, regression, unexpected behavior OR MUST USE when asks to fix, repair, investigate, debug, diagnose, apply the fix from review, trace a value, or verify that code works correctly. Fires before diagnosing - enforces tracing the entire code flow step by step instead of guessing the cause, proving the diagnosis with a failing test, then handing the proven fix plan to `simpleplan`.Votes: 0GitHub stars: 8
- Simpleplan ReviewerInvoked only by simpleplan skill.Votes: 0GitHub stars: 8
- SimpleplanWrites the implementation plan for a small, contained change already understood in context - an intent interview or a proven bug diagnosis. Invoked by the intent or simpledebug skill, or by explicit user command only - never spontaneously, never before an interview.Votes: 0GitHub stars: 8
- SuperbuildBuild orchestrator for an approved SuperPlan. Use it ONLY when the approved plan's body says to build it with the superbuild skill - never for a plan that names simplebuild, never without an approved plan.Votes: 0GitHub stars: 8
- Superdev MemoryUse ALWAYS when the user wants to create, initialize, regenerate, bootstrap, or maintain CLAUDE.md project-memory for a repository - set up project memory, add a memory layer, or make Claude understand the codebase. Triggers include "create CLAUDE.md", "initialize project memory", "bootstrap Claude context", "set up CLAUDE.md", "add memory layer". Generates a hierarchical CASCADE of CLAUDE.md files (one general root plus progressively more specific child nodes in genuine architectural units),...Votes: 0GitHub stars: 8
- Superdev RulesUse ALWAYS when the user wants to discover, capture, or maintain project coding conventions as .claude/rules files - learn the codebase's conventions, set up path-scoped rules, teach Claude the project's style. Triggers include "create rules", "set up .claude/rules", "capture coding conventions", "learn our conventions", "add naming/testing/error-handling rules", "audit rules". Discovers dominant patterns from the host code with real examples, confirms each in an interview, and writes MANY SM...Votes: 0GitHub stars: 8
- Superplan ReviewerInvoked only by superplan skill.Votes: 0GitHub stars: 8
- SuperplanWrites the implementation plan for a reviewed spec. Invoked by the superspec skill only, with the spec path as its argument - never directly, never without a reviewed spec.Votes: 0GitHub stars: 8
- Superspec RefineUse this skill always when user want to work on existing spec.Votes: 0GitHub stars: 8
- Superspec ReviewerInvoked only by superspec skill.Votes: 0GitHub stars: 8
- SuperspecInvoked by intent skill only.Votes: 0GitHub stars: 8
- TddTest-Driven Development discipline expert. Use for a plan task marked `TDD: required`, or when the user explicitly asks for test-first work. Enforces Red-Green-Refactor (iron law: no production code without a failing test first - violations are deleted and restarted), mandatory VERIFY-RED and VERIFY-GREEN checkpoints, no horizontal slicing, and a per-cycle stop-condition checklist. Triggers include "TDD", "test first", "red-green-refactor", "RGR". Do NOT use for a task marked `TDD: none`, for...Votes: 0GitHub stars: 8
- Code AuditorMulti-agent code audit through one lens (bugs, security, web-performance, runtime-performance, tests or design) over one scope (the current diff, one directory or the whole repository), every finding verified by an independent agent. Use when the user asks to audit, review or hunt defects in code for one of those lenses or scopes, such as "audit security of my branch" or "check the tests in src/api".Votes: 0GitHub stars: 8
- CommitUse whenever the user wants to commit, save, or record changes to git - including "commit", "commit changes", "commit all", "commit these files". This is the ONLY path to a commit: never run git add/commit yourself, never branch, never inspect git status/diff first - a forked agent stages, writes the Conventional Commits message, commits and verifies. Arguments, all optional - no arguments commits every change (modified, new and deleted files); one or more space-separated paths (files or dire...Votes: 0GitHub stars: 8
- Create IssueCreates a GitHub issue from the project's issue form templates. Use when the user asks to open an issue, not to fix a bug.Votes: 0GitHub stars: 8
- Create PrOpens a pull request for the current branch. Only on the user's request.Votes: 0GitHub stars: 8
- E2eGenerates and locally verifies Playwright tests for the QA scenarios of one viber run, then commits them. Use it after a build closed with a qa.e2e.md handoff file in its run directory.Votes: 0GitHub stars: 8
- ExtensionCreates or adapts an agent in the project's .claude/agents/ that a viber build runs at its close, just before the run is archived, with one forked skill in .claude/skills/ per phase when its work runs in phases one after another, and registers it as an entry of build.extensions in .claude/viber.yml. Use when the project needs a closing step of the project's own, for example end-user help files written after every task is committed.Votes: 0GitHub stars: 8
- FixerTraces a reported bug to its root cause and proves it with a failing test, then hands the diagnosis to the planner - it applies no fix itself. Use whenever user reports a bug and wants to fix it. Not for a fix the user asked to apply directly, without a plan.Votes: 0GitHub stars: 8
- HandoffWrites a handoff file that lets a fresh session continue the current conversation - where to look, what is done, the decisions with the alternatives they rejected, what comes next and the open problems. Saved under .temp/viber/handoff/ unless the first argument is a directory or .md path (it holds a `/`, ends in `.md` or is `.`); any other text is a loose prompt steering the focus.Votes: 0GitHub stars: 8
- HelpOpens the viber usage guide in the default browser.Votes: 0GitHub stars: 8
- Idea ValidatorValidates a business idea as a side-project candidate. Runs web research (problem, market, competition), scores it on 9 dimensions including Distribution, Side-project fit and Autopilot fit, convenes a 7-member council of independent perspectives, and produces a single self-contained HTML report with scorecard, dissent, experiment plan and pre-committed decision thresholds. User-invoked only.Votes: 0GitHub stars: 8
- ImplementorBuilds an approved plan task by task. Requires an existing plan; without one, suggest the viber:intent interview.Votes: 0GitHub stars: 8
- IntentPlanning interview - asks what the conversation and the code leave open, one question at a time, sizes the scope, proposes the spec shape and hands a confirmed summary to viber:planner. Never started unprompted. When a change looks like it needs a plan and no confirmed interview or viber:fixer diagnosis is in context, keep talking with the user and suggest, in one line, this interview or `plain plan mode` (Claude Code's own plan mode, no interview) for a small, well-understood change; invoke ...Votes: 0GitHub stars: 8
- MemoryReviews and repairs the host project's CLAUDE.md cascade - maps every node with its own size and the size of the chain a reader loads with it, creates the nodes a project with none needs, verifies each existing node against the code of the area it describes, brings every node below the root within budget, and resets the layer on demand. Use whenever the user wants to create, bootstrap, initialize, refresh, audit or reset project memory, or asks which nodes went stale, lost their area or grew ...Votes: 0GitHub stars: 8
- PlannerOnly for a confirmed viber:intent interview or a viber:fixer diagnosis already in context - never the entry point. Without one, suggest the viber:intent interview and let the user decide. Turns that input into a reviewed implementation plan - acceptance criteria, file map, then tasks carrying dependencies, contracts, verification and DoD.Votes: 0GitHub stars: 8
- Pro DesignerProfessional UI/UX design standards to avoid AI slop. Use when the look of an interface is the goal - a new screen or page, a redesign, a UI review or critique, choosing colors, type, spacing, layout or motion, or a UI that looks generic or AI-generated. Not for a small functional change.Votes: 0GitHub stars: 8
- PrototypeHTML mockup of a UI change in the project's own look, one proposal or three variants. Use when the user wants to see or compare a UI change before planning it.Votes: 0GitHub stars: 8
- RulesReviews and repairs the host project's .claude/rules/ directory - maps every rule with its own size, the scope it declares and the number of tracked files that scope really matches, proposes the rules a scope with none needs, verifies each existing rule against the code it gates, and resets the layer on demand. Use whenever the user wants to create, review, audit, refresh or reset the project's coding rules, or asks which rules went stale, match nothing or grew past their budget.Votes: 0GitHub stars: 8
- SetupPrepares a project for viber - seeds .claude/viber.yml, .gitignore and the recommended .claude/settings.json permissions (asking whether to merge into or reset an existing one that differs from them), checks the gh CLI is installed, checks whether CLAUDE.md names the build and test commands, and opens the usage guide in the browser.Votes: 0GitHub stars: 8
- Skill DesignerAuthoring doctrine for Claude Code skills and agents: one responsibility per skill, triggering description, imperative body, progressive disclosure, fork placement, lint, and per-model prompting profiles for Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5. Also for auditing or shrinking an existing SKILL.md or agent file, choosing `model:` or `effort:` for one, porting it to another model, or when an agent misbehaves on one model only (stops early, overreaches its scope, under-reports findings,...Votes: 0GitHub stars: 8
- TalkResearch conversation on one topic - a free exchange of views in which Claude takes and defends positions, challenges weak ones and has every claim that carries weight checked against trusted sources on the web through viber:researcher, ending on conclusions with their sources, never on code, a plan or an issue. Takes the topic or an opening thesis as its argument.Votes: 0GitHub stars: 8
- TriageReviews one reported issue against the current code - whether it can be done, how, what it affects and how big the change is - classifies it and names the next viber step without taking it. Takes a GitHub issue number or URL, or pasted issue text, and can post the report as a comment on the issue.Votes: 0GitHub stars: 8
- Idea ValidatorValidates a business idea as a side-project candidate. Runs web research (problem, market, competition), scores it on 9 dimensions including Distribution, Side-project fit and Autopilot fit, convenes a 7-member council of independent perspectives, and produces a single self-contained HTML report with scorecard, dissent, experiment plan and pre-committed decision thresholds. User-invoked only.Votes: 0GitHub stars: 6
- Model PromptingPer-model prompting profiles for Claude Code skills and agents - how Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 4.5 read instructions, where each one fails and the wording that fixes it, plus the rules for a file that runs on several models. Use when choosing `model:` or `effort:` for a skill or agent, writing or tuning one for a named model, porting one to another model, or when an agent misbehaves on one model only (stops early, overreaches its scope, under-reports findings, skips parallel c...Votes: 0GitHub stars: 6
- Skill DesignerAuthoring doctrine for Claude Code skills and agents: one responsibility per skill, triggering description, imperative body, progressive disclosure, fork placement, lint. Also for auditing or shrinking an existing SKILL.md or agent file.Votes: 0GitHub stars: 6
- Pro DesignerProfessional UI/UX design standards to avoid AI slop. ALWAYS use whenever creating, styling or reviewing ANY user interface - a page, screen, dashboard, form, onboarding or pricing flow, landing page, navigation, or a single component - even if the user only says "build/add/fix" and never says "design". Also use when critiquing existing UI, adding animations, making a static page feel alive, choosing colors, fonts, spacing, or layout, or when a UI looks generic, templated, or AI-generated.Votes: 0GitHub stars: 6
- Code AuditorPrioritized, multi-agent investigation of a large codebase using the Impact x Opportunity law.Votes: 0GitHub stars: 6
- CommitUse whenever the user wants to commit, save, or record changes to git - including "commit", "commit changes", "commit all", "commit these files". This is the ONLY path to a commit: never run git add/commit yourself, never branch, never inspect git status/diff first - a forked agent stages, writes the Conventional Commits message, commits and verifies. Arguments, all optional - no arguments commits every change (modified, new and deleted files); one or more space-separated paths (files or dire...Votes: 0GitHub stars: 6
- Create IssueCreates a GitHub issue from the project's issue form templates. Use when the user asks to open an issue, not to fix a bug.Votes: 0GitHub stars: 6
- Create PrOpens a pull request for the current branch per the project's branching - title from github.pr-title, body from the work entry's pull request template, after a preview.Votes: 0GitHub stars: 6
- E2eGenerates and locally verifies Playwright tests for the QA scenarios of one viber run, then commits them. Use it after a build closed with a qa.e2e.md handoff file in its run directory.Votes: 0GitHub stars: 6