Ecosystem Standards

Ecosystem Standards Review Protocol

This skill details how to perform an audit on new or existing capabilities (Skills, Plugins, Workflows, Sub-Agents, and Hooks) against authoritative ecosystem specifications to ensure they are created, installed, and structured correctly.

Instructions

When invoked to review a codebase component or a planned extension:

1. Identify the Component Type: Determine if the subject is a Plugin boundary, an Agent Skill, an Antigravity Workflow/Rule, a Sub-Agent, or a Hook.
2. Recall the Specs: Before reviewing, read the relevant specification file found in the ecosystem-authoritative-sources skill library.
   • Path: ../../references/*.md
3. Perform Rigorous Audit:
   • Structure: Does the directory schema match the standard (.claude-plugin/plugin.json, my-skill/SKILL.md)? Are all supporting files strictly organized into the official optional directories (scripts/, references/, assets/) rather than cluttering the skill root?
   • Manifest Schema: Does plugin.json follow the authoritative schema? Check:
     • name is kebab-case (lowercase, hyphens, no spaces)
     • version uses semver (e.g., 0.1.0, not 1.0)
     • author is an object {"name": "..."}, NOT a string
     • No author.url field (not in spec)
     • No commands_dir or skills_dir fields (auto-discovered)
     • skills, agents, hooks, commands, scripts, dependencies arrays are documentation-only (ignored by runtime, OK to keep for human readability)
     • See references/plugins.md in ecosystem-authoritative-sources for the full schema
   • Naming: Verify the skill name uses the gerund form (verb + -ing, e.g., analyzing-spreadsheets). Reject generic nouns. Ensure the name is 1-64 lowercase alphanumeric chars/hyphens only, contains NO consecutive hyphens (--), and EXACTLY matches the parent directory name.
   • Content: Does the YAML frontmatter adhere precisely to rules (description 1-1024 chars, compatibility max 500 chars, metadata strictly string-to-string keys/values)? Provide the recommendation to run skills-ref validate ./my-skill to definitively catch parse errors.
   • Description Viewpoint: Ensure the description is written strictly in the third person ("Extracts text", not "I extract text") and isn't overly vague.
   • Progressive Disclosure: For Skills, is the SKILL.md file appropriately constrained (< 500 lines) with extraneous detail pushed to one-level deep reference files? Reject deeply nested reference chains.
   • Reference Paths: Verify that all file references are strictly relative to the skill's root (e.g., scripts/extract.py), avoiding absolute paths outside the plugin boundaries.
   • Reference Readability: Do reference files >100 lines contain a Table of Contents for partial-read navigation?
   • Script Quality: Verify python utility scripts do not punt errors back to the LLM (e.g., failing silently), but instead handle exceptions safely or emit clear stderr messages. Ensure they don't use undocumented "magic numbers" (voodoo constants).
   • Multi-CLI Support: When integrating agent CLI plugins, support exists for claude-cli, gemini-cli, and copilot-cli. Plugins must reflect the native CLI syntax in their system files.
   • Anti-Patterns: Check for hardcoded credentials and Windows style paths (\).
   • Connector Abstraction: If the plugin uses MCP tools, does it include a CONNECTORS.md using the ~~category abstraction pattern instead of hardcoding specific tool names? This is required for portability. Ensure all native tool calls use the fully qualified ServerName:tool_name format.
   • Interaction Design Quality: For skills with user interaction, verify they use appropriate patterns:
     • Discovery phases use progressive questioning (broad → specific), not question walls
     • Decision points offer numbered option menus (3-7 items max)
     • Expensive operations have confirmation gates
     • Multi-step workflows include inline copyable checklists.
     • Skills end with next-action menus, not dead ends
     • Workflows taking long documents gracefully degrade using Document Format Agnosticism.
   • Execution Safety (Plan-Validate-Execute): Do destructive or massive workflows mandate an intermediate verifiable plan file (e.g., changes.json) before execution?
   • Dual-Mode Architecture: If the skill both creates new artifacts AND improves existing ones, verify it implements the Bootstrap + Iteration dual-mode pattern with separate sections and trigger phrases.
   • Output Templates: If the skill generates reports or artifacts, verify it either defines an output template or negotiates the format with the user.
   • Escalation and Safety: Workflows with external risk must explicitly implement Graduated Autonomy Routing and Escalation Trigger Taxonomies rather than blanket-stopping on all issues.
   • Source Transparency: Data synthesis output MUST conclude with explicit Sources Checked and Sources Unavailable blocks.
4. Produce Feedback: Provide explicit, granular feedback outlining exactly which ecosystem constraints were violated and concrete suggestions for fixing them. Ensure your feedback is actionable.