shield_lock
JinDaGe
← Back to skills
extension
Category: AI Agent CapabilitiesNo API key required

using-arc

Use when starting any conversation - establishes Arc's skill routing, instruction priority, and bootstrap rules

personAuthor: jakexiaohubgithub
open_in_newRepositorydownloadDownload package
<SUBAGENT-STOP> If you were dispatched as a subagent to execute a specific task, skip this skill. </SUBAGENT-STOP>

<arc_runtime> Requires the full Arc bundle. Arc-owned paths (agents/, references/, disciplines/, templates/, scripts/, rules/, skills/) resolve from the plugin root — the directory containing agents/ and skills/. Everything else is the user's repository.

Prompt-only installers (skills.sh and similar) copy SKILL.md files without Arc's bundled material. If a workflow needs Arc-owned files and only prompt-only skills are present, stop early and tell the user to upgrade to the full Claude, Codex, or Cursor plugin install. </arc_runtime>

Using Arc

Arc's canonical product definition, domain language, and operating boundary live in CONTEXT.md at the Arc repository root. This skill applies that context at session startup.

Arc has a broad workflow surface. Use this skill as the small control plane that decides how to route work without loading every workflow into context at once.

Arc is self-contained and lifecycle-focused. Its job is to move software work from idea to shipped code through Arc-owned workflows, not to act as a general registry for external personal skills. When a specialist practice is useful, absorb it into the relevant Arc workflow in Arc's own language.

Instruction Priority

When instructions conflict, use this order:

  1. User instructions in the conversation
  2. Project instructions (AGENTS.md, CLAUDE.md, repo docs)
  3. Arc skills
  4. Default system behavior

The user stays in control. Arc provides process, not authority over explicit user intent.

The Rule

Before substantial work, decide whether an Arc skill clearly applies.

  • If the user names an Arc skill, use it.
  • If the task clearly matches an Arc workflow, use that skill before acting.
  • If the task is small, direct, or outside Arc's workflows, respond normally.

Arc should improve routing, not create ceremony for every trivial request.

Do not route users to external skill collections as part of Arc's normal operation. External skills can inspire Arc workflow design, but Arc workflows must remain usable on their own.

Platform Adaptation

<required_reading> Arc skills may mention Claude Code tool names. For platform mappings and equivalents, read:

references/platform-tools.md </required_reading>

When a skill says AskUserQuestion, preserve the behavior rather than the literal tool name. In Codex, ask one concise plain-text question at a time unless a structured question tool is actually available in the current mode. In Cursor, ask one concise question at a time (structured prompts when the harness supports them; otherwise plain text). Do not narrate tool fallbacks or tell the user that a question tool is unavailable.

For UI work, keep these roles separate:

  • WireText -> low-fidelity wireframes and layout exploration
  • Chrome MCP -> preferred rendered-page verification in Claude Code
  • agent-browser -> preferred browser automation fallback outside Claude Code
  • Playwright -> scripted browser fallback when needed
  • Figma MCP -> implementation from real design files

Progressive Disclosure

Do not preload large Arc workflows.

  • Start with the smallest relevant skill
  • Load reference files only when the active task actually needs them
  • Prefer targeted rules and references over broad up-front reading

references/index.md catalogues every Arc reference by concern. Consult it when a task needs background you can't name a file for.

Workflow Routing

Use these defaults:

  • Project purpose, goals, or domain language -> vision
  • New feature or product thinking -> ideate
  • Plan execution -> implement
  • Small scoped change -> implement
  • Architecture or quality review -> review or audit
  • Testing work -> testing
  • Go-live/shareability readiness -> launch
  • Commit, push, or simple publish -> commit (cross-cutting)
  • Version bumps, changelogs, or coordinated multi-package releases -> release (cross-cutting)
  • Turn findings into an executable backlog, or reconcile existing plans -> improve (cross-cutting)
  • Structural friction, god files, or a refactoring plan -> refactor (cross-cutting)
  • Unsure what to do -> ask one focused question, then route to the smallest matching workflow

When a task needs a specialist lens, keep the routing Arc-native:

  • Ambiguous product intent -> vision, ideate, or review with one focused clarifying question
  • Reuse or duplicated UI patterns -> implement, review, or audit
  • Public API documentation -> package-focused implement or external documentation skills
  • Architecture boundaries or oversized modules -> review, audit, or implement
  • Rendered UX, browser behavior, and responsive issues -> implement when tied to Arc UI work, otherwise external rendered-app QA skills

Library track: For a single-package repo (a library or CLI rather than a deployed app), the common lifecycle is ideate (light) -> implement -> release. Vision and launch carry explicitly optional weight here — a library has no go-live URL, so reach for them only when they genuinely help.

Artifact Locations

Arc-owned artifacts live under:

  • docs/arc/specs/
  • docs/arc/plans/
  • docs/arc/plans/INDEX.md
  • docs/arc/audits/

The vision artifact is the repository's root CONTEXT.md (plus CONTEXT-MAP.md for multi-context monorepos), not a docs/ file.

Architectural decisions live in docs/adr/ — that is the canonical location. Do not create or route to docs/arc/decisions/. If a repo already has both docs/adr/ and a legacy docs/arc/decisions/, treat docs/adr/ as canonical and migrate decisions out of docs/arc/decisions/.

If a workflow references legacy docs/plans/, treat it as a compatibility fallback while the repo migrates.