Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add anatolykoptev/dozor --skill claude-escalationgit clone --depth 1 https://github.com/anatolykoptev/dozorWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/anatolykoptev/dozor/claude-escalation)<a href="https://agentmods.dev/skills/anatolykoptev/dozor/claude-escalation"><img src="https://agentmods.dev/badge/skills/anatolykoptev/dozor/claude-escalation/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/anatolykoptev/dozor/claude-escalation"><img src="https://agentmods.dev/badge/skills/anatolykoptev/dozor/claude-escalation.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00052 | $0.00577 |
| Opus 5 | $0.00026 | $0.00289 |
| Sonnet 5 | $0.00010 | $0.00115 |
| Haiku 4.5 | $0.00005 | $0.00058 |
Grade A, and why
claude-escalation scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Claude Code Escalation
When you cannot complete a task within the iteration budget, escalate to Claude Code for deep investigation.
When to Escalate to Claude Code
- You received a system warning about remaining iterations
- Same error repeats after 3+ fix attempts
- Task requires reading/editing source code
- Build or deploy failures with cryptic errors
- Configuration issues requiring code-level analysis
When NOT to Escalate
- Container just needs a restart (auto-fix handles this)
- Rate limits / quota errors (self-healing)
- Benign warnings (qdrant telemetry, searxng CAPTCHA, go-hully pool)
- Transient network errors that resolve within minutes
Escalation Budget
You have MAX 5 tool calls before you MUST escalate or give up. Do not spend iterations on repeated diagnostics.
How to Escalate
Call claude_code with async=true and a structured prompt:
claude_code(
prompt: "<full context — see template below>",
async: true
)
Always use async=true — Claude Code tasks take time and will notify you via Telegram when done.
Prompt Template
## Task
<original task from the user>
## What was done
<list tools called and their results>
<what worked, what failed>
## Current error
<exact error text from logs>
## Logs (recent entries)
<paste relevant log lines>
## What needs to be done
<specific ask: fix the bug / investigate / change config>
Example
## Task
Deploy api-service, but deploy fails at build stage.
## What was done
- server_deploy(action: deploy) → failed
- server_inspect(mode: logs, service: api-service) → compilation error
## Current error
Error: cannot find module github.com/example/core
## Logs
2026-02-18 18:20:05 ERROR build failed: missing import
2026-02-18 18:20:05 ERROR go: github.com/example/[email protected]: not found
## What needs to be done
Find the root cause and fix go.mod or add the missing dependency.
After Escalating
- Tell the user: "Task delegated to Claude Code — result will be sent separately."
- Do NOT retry the same failing operations
- When Claude Code responds via Telegram, relay the result to the user
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 89 lines · 52 tokens per session scan A 7614e47c3601
claude-escalation is a skill published in the GitHub repository anatolykoptev/dozor (5 stars, last pushed today), licensed MIT. It adds 52 tokens to every session and 577 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
run-llms
Comprehensive guide for setting up and running local LLMs using Harbor. Use when user wants to run LLMs locally, set up or troubleshoot Ollama, Open WebUI, llama.cpp, vLLM, SearXNG, Open Terminal, or similar local AI services. Covers full setup from Docker prerequisites through running models, per-service…
harbor
CLI toolkit for managing containerized LLM services. Use when the user wants to start, stop, configure, or manage AI/LLM services like Ollama, Open WebUI, llama.cpp, vLLM, LiteLLM, ComfyUI, and 250+ others. Triggers on requests to "run a model", "start ollama", "set up an LLM", "configure harbor", "manage services"…
harbor-daytona
Use Harbor's Daytona sandbox platform for computer use — creating sandboxes, taking screenshots, sending mouse/keyboard input, and building agent loops. Use when the user wants to interact with a GUI, automate a desktop, do computer use, control a browser visually, or run Claude computer use against a Daytona sandbox.
boost-modules
Skill "boost-modules" from av/harbor, covering harbor boost custom modules, module structure, quick reference, output methods and stream text to client.
query-snmp-traps
Query SNMP trap logs through Netdata Cloud or directly from a Netdata Agent. Use when the user asks about SNMP traps, trap journal entries, trap severities, trap categories, trap senders, deduplication summaries, decode errors, TRAP fields, TRAPVAR indexed varbind fields, TRAPJSON varbind audit data, or how to inspect…
agent-tool
Add a new tool/function the AI agent can call (e.g. look something up, hit an external API, perform an action). Use when extending the assistant's capabilities, wiring a new function into the agent, or when the model needs a new action. This project uses {{ cookiecutter.aiframework }}.