Sk Claude Code Conventions

Claude Code Conventions — Reference

Model-driven reasoning scaled by the effort parameter. A mechanism to persist static context (system prompts, skills) to reduce latency and cost. Structuring content into layers (Body → References) to minimize context bloat.

Model & Reasoning Selection

| Tier | Model ID | Application | | :--- | :--- | :--- | | triage | claude-haiku-4-5-20251001 | High-volume, low-latency classification. | | fast | claude-haiku-4-5-20251001 | Reviews and bounded worker steps. | | medium | claude-sonnet-4-6 | Default for pipelines, agents, and skills. | | deep | claude-opus-4-8 | High-complexity architecture or auditing tasks. |

This skill is Claude Code-scoped (Tier 1 only). Concrete model IDs above describe the Tier 1 resolution of the MODEL_SELECTION: TIER_BASED_RUNTIME_RESOLVED invariant — i.e., Tier 1's platform_profile.model_tiers.medium resolves to claude-sonnet-4-6, .fast/.triage to claude-haiku-4-5-20251001, and .deep to claude-opus-4-8. On other tiers the same fast/deep slots resolve to that platform's idiomatic models (see skills/sk-platform-dispatch/profiles/{tier_id}.json). Per DEPENDENCY_INVERSION: PROFILE_DRIVEN, pipeline-architect and other tier-agnostic skills MUST consult platform_profile.model_tiers rather than the CC literals shown here. Reasoning scales within-tier via effort; tier switches are warranted only when the step category genuinely differs.

Caching & Context Discipline

1. CACHE STABILITY

• Static First: Context ordered as System Prompt → Skills → Tools → Conversation History.
• Invariant Toolsets: Tool and skill lists must not be mutated mid-session; mutation invalidates the entire prefix.
• Static Headers: Dynamic timestamps or mutable data must not appear in the static system prefix.

2. CACHE BREAKPOINTS

• Up to 4 cache_control breakpoints are supported.
• The final breakpoint belongs on the last block that remains identical across requests.
• Agentic loops use 1-hour TTL to prevent cache expiration between turns.

Artifact Schemas

Skill Frontmatter

SKILL.md bodies must not exceed 500 lines. Descriptions must use third-person voice and focus strictly on triggering conditions.

name: {lowercase-hyphens}
description: {triggering-only, third person}
effort: {low | medium | high | xhigh | max}
disable-model-invocation: {boolean}
user-invocable: {boolean}

Agent Frontmatter

Agent bodies must not exceed 150 lines. Depth belongs in companion -references/references/*.md files.

name: {identifier}
description: {action-oriented trigger}
effort: {high for architects, medium for workers}
isolation: worktree
skills: [sk-4d-method, sk-spec-driven-development]

Plugin-scope enforcement rule: Claude Code ignores permissionMode, hooks, and mcpServers on subagents loaded from a plugin — only tools: and disallowedTools: are enforced on that path. Reviewer write-deny for plugin-shipped agents therefore rests entirely on the tool allowlist; permissionMode in agents/*.md is declarative intent, not a barrier. Project-scope agents (including agents MATERIALIZE writes under .claude/agents/superpipelines/{P}/) enforce permissionMode normally. No document may claim permissionMode enforcement for plugin-shipped agents (auditor BUNDLE-08).

Progressive Disclosure Checklist

• [ ] SKILL.md body ≤500 lines.
• [ ] References are exactly one level deep from SKILL.md.
• [ ] References >100 lines include a Table of Contents.
• [ ] Descriptions are third-person and triggering-only.
• [ ] Examples are concrete and specific to the domain.

• memory: project must not appear in agent frontmatter.
• All relative paths must be resolved via ${CLAUDE_PLUGIN_ROOT} or ${CLAUDE_PLUGIN_DATA}.
• All pipeline state must be persisted to the canonical pipeline-state.json location.

Reference Files

• sk-pipeline-state/SKILL.md — State management rules.
• sk-pipeline-paths/SKILL.md — Scope-aware path resolution.
• sk-4d-method/SKILL.md — Per-invocation wrapper.
• sk-spec-driven-development/SKILL.md — SDD protocol.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: gustavo-meilus
• Source: gustavo-meilus/superpipelines
• License: MIT

Install and usage instructions live in the source repository linked above.