Add Codex metadata and agent skill evals

.agents/plugins/marketplace.json

@@ -0,0 +1,20 @@
+{
+  "name": "apify-agent-skills",
+  "interface": {
+    "displayName": "Apify Agent Skills"
+  },
+  "plugins": [
+    {
+      "name": "apify-agent-skills",
+      "source": {
+        "source": "local",
+        "path": "./"
+      },
+      "policy": {
+        "installation": "AVAILABLE",
+        "authentication": "ON_INSTALL"
+      },
+      "category": "Data"
+    }
+  ]
+}

.codex-plugin/plugin.json

@@ -0,0 +1,30 @@
+{
+  "name": "apify-agent-skills",
+  "version": "2.0.0",
+  "description": "Official Apify agent skills for web scraping, data extraction, actor development, and automation.",
+  "author": {
+    "name": "Apify",
+    "email": "support@apify.com",
+    "url": "https://apify.com"
+  },
+  "homepage": "https://github.com/apify/agent-skills",
+  "repository": "https://github.com/apify/agent-skills",
+  "license": "Apache-2.0",
+  "keywords": ["apify", "scraping", "automation", "actors", "data-extraction", "agent-skills"],
+  "skills": "./skills/",
+  "interface": {
+    "displayName": "Apify Agent Skills",
+    "shortDescription": "Scraping and automation skills for agents.",
+    "longDescription": "Apify Agent Skills help agents scrape sites, build and deploy Actors, actorize existing code, generate output schemas, and integrate with the Apify API.",
+    "developerName": "Apify",
+    "category": "Data",
+    "capabilities": ["Search", "Automation", "Code"],
+    "websiteURL": "https://apify.com",
+    "defaultPrompt": [
+      "Scrape this market into a CSV.",
+      "Create an Apify Actor from this script.",
+      "Generate an Actor output schema."
+    ],
+    "brandColor": "#20A34E"
+  }
+}

README.md

@@ -116,6 +116,15 @@ These skills ship with `agents/AGENTS.md` (auto-generated index) and `gemini-ext
 git clone https://github.com/apify/agent-skills
 ```
 
+The repository also includes Codex plugin metadata for installing the same
+skills as a plugin:
+
+```bash
+codex plugin marketplace add https://github.com/apify/agent-skills
+```
+
+Then install **Apify Agent Skills** from the Codex Plugins UI.
+
 ### Any other agent that reads Markdown
 
 Reference the skill files directly:
@@ -149,6 +158,20 @@ Reference the skill files directly:
 
 Apify Actors use pay-per-result or pay-per-event pricing, set independently by each Actor. Free tier credits are included with every account. Check individual Actor pricing on the [Apify Store](https://apify.com/store).
 
+## Evals and production telemetry
+
+The `evals/apify-agent-skills/` directory contains a small human-review eval set
+covering the Ultimate Scraper, Actor development, and output-schema workflows.
+The cases are harness-neutral so the same plugin behavior can be checked in
+Claude Code, Codex, Gemini CLI, or another agent workspace.
+
+If you publish the plugin through Telvine, keep runtime telemetry metadata-only:
+`skill.invocation.start`, `skill.invocation.end`, and `skill.invocation.error`
+for skill behavior, plus `plugin.component.invoked` and
+`plugin.component.error` for non-skill components. Do not emit prompts, scraped
+records, API tokens, connector payloads, tool arguments, browser captures, or
+model outputs.
+
 ---
 
 ## Support

evals/apify-agent-skills/cases.jsonl

@@ -0,0 +1,3 @@
+{"id":"ultimate-scraper-leads","input":"Find Italian restaurants in Brooklyn rated under 4 stars on Google Maps, collect website URLs, and outline a CSV schema for outreach.","expected_outcome":"Uses the Ultimate Scraper skill, selects appropriate Apify Actors or fallback search, includes fields for source attribution, and avoids claiming data was collected unless an Actor run occurred."}
+{"id":"actor-development","input":"Create a minimal JavaScript Apify Actor that accepts a startUrl input, crawls one page, and pushes title plus URL to the dataset.","expected_outcome":"Uses Actor.init/exit or current SDK patterns, defines input/output expectations, handles missing input, and includes a runnable file structure."}
+{"id":"output-schema","input":"Given actor output objects with title, url, price, rating, and scrapedAt, generate an Apify dataset schema.","expected_outcome":"Uses the output schema skill, assigns appropriate field types, marks required fields thoughtfully, and keeps schema compatible with Apify conventions."}

evals/apify-agent-skills/rubric.md

@@ -0,0 +1,27 @@
+# Apify Agent Skills eval rubric
+
+Score each case from 1-5.
+
+## Skill routing
+
+- 5: Selects the right Apify skill for scraping, actor development, actorization, schema generation, or SDK integration.
+- 3: Uses Apify concepts but needs extra prompting to choose the right path.
+- 1: Ignores the available Apify skills.
+
+## Apify correctness
+
+- 5: Uses current Apify CLI, SDK, Actor, dataset, and schema conventions.
+- 3: Mostly correct with minor naming or setup issues.
+- 1: Suggests APIs or files that do not fit Apify.
+
+## Output quality
+
+- 5: Produces implementation-ready steps, schemas, or data plans with clear assumptions.
+- 3: Provides useful but incomplete guidance.
+- 1: Produces vague scraping or automation advice.
+
+## Privacy and data boundaries
+
+- 5: Avoids collecting credentials, private data, prompts, tool arguments, or model outputs beyond the requested task.
+- 3: Includes unnecessary data fields without sensitive content.
+- 1: Suggests unsafe scraping, secret handling, or private data exposure.