Imported from openlayer-ai/openlayer-skills (AGENTS.md). Install upstream with npx skills add openlayer-ai/openlayer-skills. Copyright stays with the author.

Agent Instructions

Guidance for agents editing this repo (adding or improving skill content).

Adding or Improving a Use Case

• Only add a use case if it beats the docs. If an agent can already serve the user by fetching the Openlayer docs (docs.openlayer.com, see references/docs-access.md), add nothing. Reserve new content for where docs fall short and the agent needs extra context (mode routing, the MCP-vs-SDK-vs-REST decision, non-obvious pitfalls). Every addition is maintenance surface and dilutes the skill.

• You should almost never touch the top-level description in SKILL.md. It only controls whether the skill is invoked, and a user asking about Openlayer, tracing, evals, or openlayer.json already triggers it. Keep it short; in-skill routing handles the rest. (It must stay under 1024 characters and contain no < or >.)

• Put "when to use" guidance in exactly two places: the one-line entry in the use-case router table in SKILL.md, and the description in the reference file's frontmatter. Nowhere else — no prose routing section in a reference body. A reference body is read only after the agent already chose to open it, so routing text there is dead weight.

• Never commit code that can go stale. Link to the relevant Openlayer docs page (append .md to any page URL) so the agent fetches current code; use short pseudo-code only for logic-specific bits. Openlayer SDKs/CLI/MCP change frequently. Committed code goes stale and teaches agents the wrong thing.

• In a reference file, less is more. Add only what's useful or what an agent couldn't infer on its own. Each reference ends with a Common Mistakes table — that is usually the highest-value part.

• Be cautious with allowed-tools. A tool not in the list still works — the user just grants permission the first time. Only auto-allow no-brainer, low-risk commands. A risky auto-allow makes people hesitate to install the skill at all.

Plugin Version Bumps

The repo ships to two marketplaces, each with its own manifest:

• .claude-plugin/plugin.json (Claude Code) and .claude-plugin/marketplace.json
• .cursor-plugin/plugin.json (Cursor)

Keep the version fields in lockstep — bump them together in the same PR (semver: patch for fixes/clarifications, minor for new capability, major for breaking renames/removals).

Validation

Run the format check before committing:

python3 scripts/validate_skills.py

This is the same check CI runs (frontmatter, cross-links resolve, no orphan references, version lockstep).