Anth Observability

What it does

When your agent uses it

• With phrases like anthropic monitoring
• Claude observability
• Anthropic metrics
• Track claude usage

Example prompts

• “anthropic monitoring”
• “claude observability”
• “anthropic metrics”
• “/anth-observability”

Requirements

• Python 3
• Compatibility (from SKILL.md): Designed for Claude Code
• Pre-approved tools (allowed-tools): Read, Write, Edit, Grep

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

1. 1Instrument the request boundary with request ID, model, status, stop reason, token aggregates, cache counters, and duration while…
2. 2Emit success and failure metrics for authentication, 4xx/5xx, 429, timeout, latency, spend, and remaining rate-limit headroom. Validate…
3. 3Test dashboards and alerts with synthetic success, timeout, rate-limit, permission, and malformed-response fixtures. Verify the alert path…
4. 4Reconcile usage through the approved authenticated server-side API on a bounded schedule, compare aggregate totals, and alert on…
5. 5Canary telemetry changes, then promote with owner approval. If redaction, cardinality, or retention checks fail, disable the new sink…

What it can do on your machine

Read from SKILL.md and the folder at commit 80f86df. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves these tools, so the agent can use them without asking each time:

  • Read
  • Write
  • Edit
  • Grep

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  No scripts in the folder and no shell commands in SKILL.md (its code samples are python).

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  Hosts in commands or code, which the agent is likely to contact:

  • api.anthropic.com

  Also links to:

  • platform.claude.com
  • status.anthropic.com

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

• Compatibility

  Designed for Claude Code

  From compatibility in the SKILL.md frontmatter.

Context cost

Anth Observability loads about 1.8k tokens when it runs. Until then it costs about 71 tokens; SKILL.md has 374 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~71

When it runs · the whole SKILL.md, loaded when a task matches

~1.8k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from jeremylongshore/tons-of-skills-marketplace at commit 80f86df, republished under its MIT licence (© jeremylongshore). 374 words, ~1,766 tokens.

name

anth-observability

description

Set up observability for Claude API integrations with metrics, logging, and alerting for latency, cost, errors, and token usage. Trigger with phrases like "anthropic monitoring", "claude observability", "anthropic metrics", "track claude usage", "claude dashboard".

allowed-tools

Read, Write, Edit, Grep

compatibility

Designed for Claude Code

version

1.7.0

license

MIT

author

Jeremy Longshore <jeremy@intentsolutions.io>

tags

saas, ai, anthropic

import anthropic
import logging
import time
import json

logger = logging.getLogger("claude")

def create_with_logging(client: anthropic.Anthropic, **kwargs) -> anthropic.types.Message:
    start = time.monotonic()
    request_meta = {
        "model": kwargs.get("model"),
        "max_tokens": kwargs.get("max_tokens"),
        "tool_count": len(kwargs.get("tools", [])),
        "stream": kwargs.get("stream", False),
    }

    try:
        response = client.messages.create(**kwargs)
        duration_ms = int((time.monotonic() - start) * 1000)

        logger.info(json.dumps({
            "event": "claude.request",
            "request_id": response._request_id,
            "model": response.model,
            "input_tokens": response.usage.input_tokens,
            "output_tokens": response.usage.output_tokens,
            "cache_read_tokens": getattr(response.usage, "cache_read_input_tokens", 0),
            "stop_reason": response.stop_reason,
            "duration_ms": duration_ms,
            "content_blocks": len(response.content),
        }))
        return response

    except anthropic.APIStatusError as e:
        duration_ms = int((time.monotonic() - start) * 1000)
        logger.error(json.dumps({
            "event": "claude.error",
            "status": e.status_code,
            "error_type": getattr(e, "type", "unknown"),
            "duration_ms": duration_ms,
            "request_id": e.response.headers.get("request-id", "unknown"),
        }))
        raise

from prometheus_client import Counter, Histogram, Gauge

claude_requests = Counter(
    "claude_requests_total", "Total Claude API requests",
    ["model", "stop_reason", "status"]
)
claude_latency = Histogram(
    "claude_latency_seconds", "Claude API latency",
    ["model"], buckets=[0.5, 1, 2, 5, 10, 30, 60]
)
claude_tokens = Counter(
    "claude_tokens_total", "Token usage",
    ["model", "direction"]  # direction: input|output|cache_read
)
claude_cost = Counter(
    "claude_cost_usd", "Estimated cost in USD",
    ["model"]
)
claude_rate_limit_remaining = Gauge(
    "claude_rate_limit_remaining", "Remaining rate limit",
    ["dimension"]  # dimension: requests|tokens
)

def track_metrics(response, duration: float):
    model = response.model
    claude_requests.labels(model=model, stop_reason=response.stop_reason, status="ok").inc()
    claude_latency.labels(model=model).observe(duration)
    claude_tokens.labels(model=model, direction="input").inc(response.usage.input_tokens)
    claude_tokens.labels(model=model, direction="output").inc(response.usage.output_tokens)

    # Cost estimation
    pricing = {"claude-haiku-4-20250514": (0.80, 4.0), "claude-sonnet-4-20250514": (3.0, 15.0)}
    rates = pricing.get(model, (3.0, 15.0))
    cost = (response.usage.input_tokens * rates[0] + response.usage.output_tokens * rates[1]) / 1e6
    claude_cost.labels(model=model).inc(cost)

# Anthropic's Usage & Cost API for billing reconciliation
# GET https://api.anthropic.com/v1/usage
# Returns daily token usage and cost per model

© jeremylongshore, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in skills/.curated/anth-observability of jeremylongshore/tons-of-skills-marketplace.

Open the folder on GitHubat commit 80f86df