mirror of
https://github.com/sentimony/skills.git
synced 2026-09-19 07:12:05 +08:00
Ihor Orlovskyi
74513851d7
* fix(git-worktree-isolation): canonicalize workspace_root and drop stale helper wording * fix(review-request): use one UNASSESSED literal across all verdicts * docs(web-debug): align heading with skill name and gate Playwright install * fix(branch-finish): separate cleanup status from outcome labels * docs(parallel-agents): drop agent-count anchor next to the no-hardcoding rule * docs(typescript): mark Python optional for helper scripts * docs(vitest): mark Python optional and name the tdd boundary * docs(echarts): use project package manager and name the debugging boundary * fix(commit-all): detect the default branch instead of assuming main or master * docs(frontend-crafting): name the debugging boundary for unexpected UI behavior * docs(tdd): name verification-gate as completion owner and narrow the trigger * feat(debugging): record symptom provenance explicitly in step 0 * docs(scope-triage): verify compatibility in the repo before asking in Route B * refactor: narrow trigger boundaries for dashfix, negafix, and maintaining-agent-context * fix(skill-crafting): require a should-not-trigger case and shorten the description * feat(plan-dev): report plan progress as a counted status line * docs: record the small-fixes release across skills * docs: adopt the single-table skill index and record the 1.39.0 release Take README.md from docs/readme-ascii-workflow, which merges the four skill tables into one and reorders the install commands, then re-apply the version bumps for the nineteen skills changed on this branch and set their Release column to v1.39.0. * docs: describe skill-crafting by what the skill does The README row carried the eval package's summary instead of the skill's own purpose; align it with the frontmatter description. * revert(skill-crafting): restore the measured description The trigger-eval run of 2026-09-16 shows the shortened description loses four positive cases that the previous one routed correctly (7/8 to 4/8 on cases clean in both arms), with no reduction in false positives. Per the plan's acceptance criterion a drop rolls the edit back, so line 3 returns to the measured baseline. The should-not-trigger requirement from the same task is behavior the run did not contest and stays in place, so 1.1.1 keeps a real package change. The skill and root changelogs drop their claims about the shortening. * docs: date the 1.39.0 release entry to its release day
Agent Skills
A collection of agent skills for Claude Code, Codex and other AI coding agents.
Install
# All at once
npx skills add sentimony/skills -a codex claude-code -y
# Or each separately
npx skills add sentimony/skills -s web-debug -a codex claude-code -y
npx skills add sentimony/skills -s vitest -a codex claude-code -y
npx skills add sentimony/skills -s typescript -a codex claude-code -y
npx skills add sentimony/skills -s echarts -a codex claude-code -y
npx skills add sentimony/skills -s scope-triage -a codex claude-code -y
npx skills add sentimony/skills -s plan-crafting -a codex claude-code -y
npx skills add sentimony/skills -s dashfix -a codex claude-code -y
npx skills add sentimony/skills -s negafix -a codex claude-code -y
npx skills add sentimony/skills -s commit-all -a codex claude-code -y
npx skills add sentimony/skills -s maintaining-agent-context -a codex claude-code -y
npx skills add sentimony/skills -s frontend-crafting -a codex claude-code -y
npx skills add sentimony/skills -s tdd -a codex claude-code -y
npx skills add sentimony/skills -s debugging -a codex claude-code -y
npx skills add sentimony/skills -s review-request -a codex claude-code -y
npx skills add sentimony/skills -s review-resolution -a codex claude-code -y
npx skills add sentimony/skills -s verification-gate -a codex claude-code -y
npx skills add sentimony/skills -s git-worktree-isolation -a codex claude-code -y
npx skills add sentimony/skills -s parallel-agents -a codex claude-code -y
npx skills add sentimony/skills -s inline-plan-dev -a codex claude-code -y
npx skills add sentimony/skills -s subagent-plan-dev -a codex claude-code -y
npx skills add sentimony/skills -s branch-finish -a codex claude-code -y
npx skills add sentimony/skills -s skill-crafting -a codex claude-code -y
Skills
| Skill | Skill Version | Release | Description |
|---|---|---|---|
| web-debug | 1.3.4 | v1.39.0 | Debug and verify local web apps via Playwright. |
| vitest | 1.3.2 | v1.39.0 | Configure, write, debug, run, migrate, and audit Vitest tests for JavaScript/TypeScript projects. |
| typescript | 1.4.2 | v1.39.0 | Configure tsconfig, diagnose compiler behavior, and audit or migrate TypeScript projects. |
| echarts | 1.2.2 | v1.39.0 | Build, audit, style, debug, and optimize Apache ECharts visualizations in vanilla JS, React, or Vue. |
| scope-triage | 1.0.6 | v1.39.0 | Classify request scope before design work, then route to direct implementation, a light spec, or a full design cycle. |
| plan-crafting | 1.3.3 | v1.34.2 | Turn an approved design or settled requirements into a bite-sized, TDD-oriented implementation plan. |
| dashfix | 1.2.3 | v1.39.0 | Ban typographic dashes in English text, check their form where a language's orthography requires them, audit a project, and score it 0-100. |
| negafix | 1.2.3 | v1.39.0 | Ban negative parallelism ("it's not just X, it's Y"), audit prose for it, and score it 0-100. |
| commit-all | 1.1.2 | v1.39.0 | Gather the working tree into a single commit on the current branch, on explicit /commit-all invocation only. |
| maintaining-agent-context | 1.3.2 | v1.39.0 | Audit, restructure, and maintain a repository's agent instruction architecture for Claude Code and Codex. |
| frontend-crafting | 1.3.2 | v1.39.0 | Create, redesign, review, and polish user interfaces with subject-driven design decisions and a verifiable quality gate. |
| tdd | 1.0.5 | v1.39.0 | Drive behavior changes through valid RED, sufficient GREEN, and evidence-backed refactoring. |
| debugging | 1.0.10 | v1.39.0 | Investigate bugs and unexpected technical behavior with a root-cause-first evidence workflow. |
| review-request | 1.0.4 | v1.39.0 | Prepare and dispatch independent code review against requirements, exact scope, and the actual diff. |
| review-resolution | 1.0.3 | v1.28.0 | Validate and resolve code-review findings with evidence, explicit dispositions, and proportional re-review decisions. |
| verification-gate | 1.0.3 | v1.29.0 | Turn a completion claim into an evidence-backed verdict against the current tree. |
| git-worktree-isolation | 1.2.1 | v1.39.0 | Select, detect, or create a safe isolated development workspace with explicit ownership, baseline, and handoff. |
| parallel-agents | 1.0.4 | v1.39.0 | Prove work units independent, isolate mutable state, and dispatch one bounded parallel wave with reconciled results. |
| inline-plan-dev | 1.0.4 | v1.39.0 | Execute an existing implementation plan inline in the current session, with plan-reality reconciliation, proportional verification, and durable resume. |
| subagent-plan-dev | 1.0.4 | v1.39.0 | Execute an existing implementation plan through scoped subagents with risk-based dispatch, independent verification, and controlled escalation. |
| branch-finish | 1.0.5 | v1.39.0 | Decide, execute and report the integration outcome for verified work, cleaning up only what is provably safe to remove. |
| skill-crafting | 1.1.1 | v1.39.0 | Create, improve, evaluate, and optimize agent skills with measured trigger boundaries and package-owned evals. |
Development Workflow
scope-triage is the development entry point. It classifies the request and picks the
minimum appropriate process, so a one-line configuration change and a new subsystem do not
receive the same ceremony.
scope-triage
│
┌──────────────┼──────────────┐
│ │ │
Route A Route B Route C
direct work light spec full design
│ │ │
│ direct work plan-crafting
│ │ │
│ │ ┌─────┴─────┐
│ │ │ │
│ │ inline-plan subagent-plan
│ │ -dev -dev
└──────────────┴────────┬───────────┘
│
implementation
│
behavior change → tdd
│
review-request (when warranted)
│
review-resolution (when findings exist)
│
verification-gate
│
branch-finish (when lifecycle exists)
│
complete
When work fails unexpectedly, the diagnostic path runs alongside the main flow and returns to whatever workflow discovered the failure.
any development stage
│
▼
debugging
│
├── browser evidence needed
│ │
│ web-debug
│ │
└──────────┘
│
root cause
│
├── material redesign needed → scope-triage
│
├── behavior correction → tdd
│
▼
fix
│
▼
return to calling workflow
Four properties of this workflow are worth stating explicitly.
tddis used within implementation, whenever behavior changes. It is a micro-cycle inside a task rather than a sequential phase that follows implementation. Both execution modes and direct implementation invoke it at the moment a behavior change is written.- Review is proportional; verification is authoritative.
review-requestis invoked when the change warrants independent eyes, andreview-resolutionvalidates findings before anything is changed in response. A review verdict stands separate from completion:verification-gateowns the authoritative pass or fail against the current tree. - The diagnostic path has three distinct owners.
debuggingowns causal investigation,web-debugsupplies browser and runtime evidence when the symptom lives in a page, andtdddrives the regression behavior once the cause is known. The fix then returns to the workflow that discovered the failure. - The workflow is composable rather than mandatory-linear. Every step after
scope-triageis reached on a condition. Route A terminates in implementation with proportional verification, andbranch-finishruns only when a branch or workspace lifecycle exists.
Conditional capabilities, invoked when the work calls for them rather than in sequence:
skill-crafting, tdd, debugging, web-debug, git-worktree-isolation, parallel-agents,
frontend-crafting, vitest, typescript, echarts.
Have fun ;)