Skip to content

2001105254-netizen/RacePulsePH / .agents/skills/racepulse-skill-maker

racepulse-skill-maker

Verified

Create, extend, register and validate RacePulsePH agent skills so they follow AGENTS.md, the frontend rule file, the antislop R-01 to R-40 rules and the skills-library.json registry. Use when adding or revising a skill, or turning a checklist or screenshot into one.

RacePulse skill maker

Every skill in this repository is registered in .agents/skills/skills-library.json and checked by a read-only validator. Author skills through this workflow so they inherit the project rules instead of drifting from them.

Rule precedence: AGENTS.md, then .agents/rules/frontend.md, then the antislop core (R-01 to R-40, C-1 to C-5), then repo skills. Details and consequences for skill content: rules and implications.

Before writing

  1. Read skills-library.json. Search skills and backlog for the topic. Extend an existing repo skill or promote a backlog entry instead of adding a near duplicate.
  2. Treat entries with origin: "vendor" (the antislop family) as read-only. Put repo-specific additions in a racepulse- skill that lists the vendor skill in dependsOn.
  3. Decide the single concern the skill owns and write down what it does not cover.

Author

  • Folder .agents/skills/<name>/. The name is lowercase, hyphenated, starts with racepulse-, and equals the frontmatter name.
  • Frontmatter holds name, description, and allowed-tools only when the skill runs scripts. The description states what the skill does, then "Use when ...", in 20 to 400 characters.
  • Keep SKILL.md under 120 lines. Move detail into references/*.md, one level deep, linked with relative paths.
  • Write imperatives. Cite real repo paths and confirm each exists with Glob before naming it.
  • Give every checklist item an action, a one-line reason (C-1) and the evidence that proves it (R-35). When a check cannot be automated, say what to inspect by hand.
  • State thresholds and claims only with a source in the repo or its docs. Otherwise write "verify against current vendor docs" (R-36, C-5).
  • Cite core rules as R-XX or C-X. Do not copy, renumber or reword them.
  • Never use the em dash character (U+2014) in any skill text (R-02).
  • Instruct agents to edit source files directly. Scripts shipped inside a skill read and report only (R-33).
  • A UI-touching skill loads with the antislop core, follows its resolved mode, and requires design direction before building (R-37).
  • A security or permission skill describes checks. It never relaxes the enforcement defined in racepulse-access.

Start from skill template.

Harvesting from lists, screenshots or articles

  1. Transcribe items verbatim with their numbers. Mark unreadable or cut-off items as gaps. Do not guess them.
  2. Remove duplicates and compare against existing skills for overlap.
  3. Cluster by concern. One skill holds one concern, usually 8 to 20 items.
  4. Map each item to this stack: Firebase Hosting, Firestore and rules, Cloud Functions, Storage, PayMongo, the RFID bridge and station. Items that do not apply, such as Kubernetes or Helm, go to deferredTopics with a reason.
  5. Turn each kept item into a check: action, reason, evidence.
  6. Record the cluster in backlog with its sources, then author it as a skill when promoted.

Register and mirror

  1. Add or update the entry in skills-library.json. Field rules: library schema. Copy the description text exactly from the frontmatter.
  2. Remove the matching backlog entry when a skill is promoted.
  3. Copy the whole skill folder to skills/<name>/ so both trees stay identical, and set mirror to that path.

Validate and deliver

Run from the repository root:

node .agents/skills/racepulse-skill-maker/scripts/validate-skill-library.cjs

Exit 0 is clean, 1 means a check failed, 2 means a usage or registry load error. Add --json for machine output. Fix failures and re-run until the exit code is 0.

Deliver with a PASS/FAIL report:

  • Interactive verification: every link opened, every documented command run once, results listed.
  • CLI/AST guard contract: validator exit code.
  • Purpose log: one line of functional justification for each section and structural choice.
brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).Stars: 322kUpdated Claude CodeCodexVerifiedInstall
brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.Stars: 197kUpdated Claude CodeCodexVerifiedInstall
frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults.Stars: 197kUpdated Claude CodeCodexVerifiedInstall
mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).Stars: 197kUpdated Claude CodeCodexVerifiedInstall