addrbookFor publishers
← The evidence ledger

Evidence row 1479262

Observed , by probe.mcp/1.

Kind
mcp.tools
About
endpoint a2810b73-cfe0-4b26-ab95-fbe6ed1e0e66
Recorded
Previous row
None: the first of its kind about this subject
Body SHA-256
091d5765a50409d325f0dff808b9dd966eb60b508365695270016eea23896858
Leaf
0x8569d40d99619c5e7ba5476385b7eef21c24c2ad33cbd49ae4c7a7611896e0a1

What was observed

{
  "count": 4,
  "tools": [
    {
      "description": "Check an agent skill (SKILL.md with its scripts and references) or AGENTS.md, CLAUDE.md, .cursorrules before loading it: prompt injection, curl | bash, encoded commands, credential theft, exfiltration, hardcoded keys, persistence. Paste the files or give one URL. Costs US$0.005 in USDC on Base (x402); nothing is executed.",
      "inputSchema": {
        "properties": {
          "files": {
            "description": "The skill files, e.g. [{path: 'SKILL.md', content: '...'}] (max 50, 256 KB)",
            "items": {
              "properties": {
                "content": {
                  "type": "string"
                },
                "path": {
                  "type": "string"
                }
              },
              "required": [
                "path",
                "content"
              ],
              "type": "object"
            },
            "type": "array"
          },
          "pinned": {
            "additionalProperties": {
              "type": "string"
            },
            "description": "Path to fingerprint from a previous audit",
            "type": "object"
          },
          "url": {
            "description": "Or a public URL of one raw text file",
            "format": "uri",
            "type": "string"
          }
        },
        "type": "object"
      },
      "name": "audit_agent_skill"
    },
    {
      "description": "Check an MCP server for tool poisoning before you connect: hidden instructions, secrecy, sensitive-file access, cross-tool shadowing and rug pulls. Pass its URL (we list its tools read-only) or paste its tools. Optional pinned fingerprints detect changed tools. Costs US$0.01 in USDC on Base (x402); servers that cannot be listed are not charged.",
      "inputSchema": {
        "properties": {
          "instructions": {
            "description": "Server instructions, if pasting",
            "type": "string"
          },
          "pinned": {
            "additionalProperties": {
              "type": "string"
            },
            "description": "Tool name to fingerprint from a previous audit",
            "type": "object"
          },
          "tools": {
            "description": "Or the tools/list result, pasted",
            "items": {
              "type": "object"
            },
            "type": "array"
          },
          "url": {
            "description": "Streamable HTTP endpoint of the MCP server",
            "format": "uri",
            "type": "string"
          }
        },
        "type": "object"
      },
      "name": "audit_mcp_server"
    },
    {
      "description": "Scan HTML or text you already have (an email, a document, a tool result) for prompt injection. Returns the risk report and the cleaned text wrapped as untrusted data. Costs US$0.002 in USDC on Base (x402). Max 256 KB.",
      "inputSchema": {
        "properties": {
          "content": {
            "description": "The HTML or text to scan",
            "type": "string"
          },
          "format": {
            "enum": [
              "html",
              "text"
            ],
            "type": "string"
          },
          "return_text": {
            "description": "Include the cleaned text (default true)",
            "type": "boolean"
          },
          "url": {
            "description": "Where the content came from (optional)",
            "type": "string"
          }
        },
        "required": [
          "content"
        ],
        "type": "object"
      },
      "name": "check_prompt_injection"
    },
    {
      "description": "Fetch a public web page or PDF and return it as clean markdown with hidden text, invisible characters and data-exfiltration links removed, plus a prompt-injection risk report. Costs US$0.005 in USDC on Base (x402); pages that cannot be read are not charged. Respects robots.txt.",
      "inputSchema": {
        "properties": {
          "format": {
            "description": "markdown keeps links; text drops them",
            "enum": [
              "markdown",
              "text"
            ],
            "type": "string"
          },
          "url": {
            "description": "Public http(s) URL",
            "format": "uri",
            "type": "string"
          }
        },
        "required": [
          "url"
        ],
        "type": "object"
      },
      "name": "read_url"
    }
  ]
}

What the leaf hashes

{"body":"091d5765a50409d325f0dff808b9dd966eb60b508365695270016eea23896858","id":1479262,"kind":"mcp.tools","method":"probe.mcp/1","observed_us":1793433867715974,"prev":null,"subject":"a2810b73-cfe0-4b26-ab95-fbe6ed1e0e66","subject_kind":"endpoint"}

Proof

Its day has not closed yet. A day closes ten minutes after midnight UTC; the proof appears here then.

How to check it · This row as JSON