AI Newsletter Digest improvements: fixed QP soft line break decoding, URL extraction, and content cleaning

archive/inactive-skills/captChas-openclaw/SKILL.md

@@ -0,0 +1,131 @@
+---
+name: captchas-openclaw
+description: OpenClaw integration guidance for CAPTCHAS Agent API, including OpenResponses tool schemas and plugin tool registration.
+homepage: https://captchas.co
+metadata: {"openclaw":{"emoji":"🧩","requires":{"env":["CAPTCHAS_API_KEY","CAPTCHAS_ENDPOINT"]},"primaryEnv":"CAPTCHAS_API_KEY"}}
+---
+
+# CAPTCHAS + OpenClaw
+
+Use this skill when integrating CAPTCHAS with OpenClaw via OpenResponses tools or OpenClaw plugin tools.
+
+## Configuration
+
+Set environment variables:
+
+- `CAPTCHAS_ENDPOINT` = `https://agent.captchas.co`
+- `CAPTCHAS_API_KEY` = `<your-api-key>`
+
+Headers:
+
+- `x-api-key`: required (use `CAPTCHAS_API_KEY`).
+- `x-domain`: optional; validated if provided.
+
+Notes:
+
+- `site_key` is optional; if omitted, it resolves from the API key or account default.
+- Avoid sending PII in `signals`.
+
+## OpenResponses Tool Schemas (OpenClaw Gateway)
+
+Use the OpenClaw `tools` array shape when calling the Gateway `/v1/responses` endpoint.
+
+```json
+{
+  "tools": [
+    {
+      "type": "function",
+      "function": {
+        "name": "captchas_agent_verify",
+        "description": "Run CAPTCHAS Agent Verify and return a decision (allow|deny|challenge).",
+        "parameters": {
+          "type": "object",
+          "properties": {
+            "site_key": {"type": "string"},
+            "action": {"type": "string"},
+            "signals": {"type": "object", "additionalProperties": true},
+            "capabilities": {
+              "oneOf": [
+                {"type": "object", "additionalProperties": true},
+                {"type": "array", "items": {"type": "string"}}
+              ]
+            },
+            "verification_mode": {"type": "string", "enum": ["backend_linked", "agent_only"]},
+            "challenge_source": {"type": "string", "enum": ["bank", "ai_generated"]},
+            "input_type": {"type": "string", "enum": ["choice", "image", "behavioral"]},
+            "media_url": {"type": "string"},
+            "media_type": {"type": "string"}
+          },
+          "required": [],
+          "additionalProperties": false
+        }
+      }
+    },
+    {
+      "type": "function",
+      "function": {
+        "name": "captchas_agent_challenge_complete",
+        "description": "Complete a challenge and mint a verification token when passed.",
+        "parameters": {
+          "type": "object",
+          "properties": {
+            "challenge_id": {"type": "string"},
+            "site_key": {"type": "string"},
+            "answer": {"type": "string"}
+          },
+          "required": ["challenge_id", "answer"],
+          "additionalProperties": false
+        }
+      }
+    },
+    {
+      "type": "function",
+      "function": {
+        "name": "captchas_agent_token_verify",
+        "description": "Verify an opaque CAPTCHAS token before completing a sensitive action.",
+        "parameters": {
+          "type": "object",
+          "properties": {
+            "token": {"type": "string"},
+            "site_key": {"type": "string"},
+            "domain": {"type": "string"}
+          },
+          "required": ["token"],
+          "additionalProperties": false
+        }
+      }
+    }
+  ]
+}
+```
+
+## OpenClaw Plugin Tool Registration
+
+Register tools using `api.registerTool(...)` and the same JSON Schema parameters as above.
+
+Example:
+
+```js
+api.registerTool({
+  name: "captchas_agent_verify",
+  description: "Run CAPTCHAS Agent Verify and return a decision (allow|deny|challenge).",
+  parameters: {
+    type: "object",
+    properties: {
+      site_key: { type: "string" },
+      action: { type: "string" },
+      signals: { type: "object", additionalProperties: true }
+    },
+    required: [],
+    additionalProperties: false
+  },
+  async execute(_id, params) {
+    return { content: [{ type: "text", text: JSON.stringify(params) }] };
+  }
+});
+```
+
+## References
+
+- Use `/v1/agent/verify`, `/v1/agent/challenge/:id/complete`, and `/v1/agent/token-verify` as the canonical API calls.
+- See `captchas-human-verification/SKILL.md` for workflow guidance.