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 agentmods add skills/kevinzai/commander/ccc-nightwatchnpx skills add KevinZai/commander --skill ccc-nightwatchgit clone --depth 1 https://github.com/KevinZai/commanderWrote 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/kevinzai/commander/ccc-nightwatch)<a href="https://agentmods.dev/skills/kevinzai/commander/ccc-nightwatch"><img src="https://agentmods.dev/badge/skills/kevinzai/commander/ccc-nightwatch.svg" alt="Measured on agentmods" 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 | $0.00043 | $0.01556 |
| Opus 5 | $0.00022 | $0.00778 |
| Sonnet 5 | $0.00009 | $0.00311 |
| Haiku 4.5 | $0.00004 | $0.00156 |
Grade D, and why
ccc-nightwatch scanned grade D with 3 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
cat ~/.claude/settings.json | grep nightwatch Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
| 🔴 Always-deny | rm -rf, git push --force to main, DROP TABLE, secret exfil patterns | Hard block + immediate alert | Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -s "https://api.telegram.org/bot$CCC_NIGHTWATCH_TG_TOKEN/sendMessage" \ How it starts
The opening of the file, as written. The whole thing — 161 lines — stays where its author put it; the contents beside it link to each section on GitHub.
$ccc-nightwatch — Remote YOLO Control
Run Claude Code autonomously overnight. Approve or deny tool use from your phone via Telegram or Discord. A PreToolUse hook intercepts every tool call, classifies it by risk, and either auto-allows, asks you, or hard-blocks — no sitting at the terminal required.
What it does
Nightwatch installs a PreToolUse permission hook that intercepts tool calls during /ccc night (YOLO mode). Each call is scored against the risk matrix. Low-risk ops (reads, searches) proceed silently. Medium-risk ops (file writes, installs) send a phone notification and wait for your tap. High-risk ops (destructive commands, force pushes) are hard-blocked and alert you immediately. Your response routes back into the session within 2–5 seconds, and Claude continues.
Triggers
$ccc-nightwatch- "I want to run Claude Code overnight"
- "approve from phone"
- "remote YOLO"
- "autonomous with safety leash"
- "let it run, I'll check from my phone"
Risk Classification Matrix
| Risk Level | Examples | Action |
|---|---|---|
| 🟢 Auto-allow | Read, Glob, Grep, LS, cat, git status, git log | Proceed silently — logged |
| 🟡 Ask human | Edit, Write, Bash (npm/yarn/pnpm install), git commit, git push (non-main) | Push phone notify → await tap → resume |
| 🔴 Always-deny | rm -rf, git push --force to main, DROP TABLE, secret exfil patterns | Hard block + immediate alert |
Unclassified tools default to 🟡 Ask human.
Architecture
Claude Code (YOLO mode)
|
[PreToolUse hook] ← hooks/nightwatch-relay.js
|
[Risk Classifier]
|
┌────┴────────────────────┐
| |
🟢 Auto-allow 🟡 Ask human 🔴 Always-deny
| | |
proceed silently [Notifier] block + alert
| / \
| Telegram Discord
| bot webhook
| \ /
| user taps ✅/❌
| |
└──────────────[Resume session]
|
tool call proceeds
(or blocked on ❌)
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.
- 6d ago First seen · 161 lines · 43 tokens per session scan D 967f1efc48d8
ccc-nightwatch is a skill published in the GitHub repository KevinZai/commander (6 stars, last pushed 2d ago), licensed MIT. It adds 43 tokens to every session and 1,556 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it D with 3 findings (reads agent configuration directories, recursive force delete, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
multi-plan-handoff
Auto-detect plan files in conversation context and pass them to multi: execute subagents by reference, not by paraphrase.
skill-dev
Use this skill when creating or refining Claude Code skill definitions. Skills are model-invoked capabilities that Claude activates autonomously based on context. Helps design focused skills with discovery-optimized descriptions, proper directory structures, and supporting resources. Automatically invoked when user…
command-dev
Use this skill when creating or refining custom Claude Code slash commands. Slash commands are user-invoked reusable prompts that can accept arguments, reference files, and execute bash operations. Helps design command syntax, argument handling, file references, bash execution, and frontmatter configuration.…
mcp-tool-dev
Create MCP tools — individual tool functions exposed via Model Context Protocol. Use this skill whenever users mention MCP tools, tool handlers, tool functions, tool definitions, or want to add capabilities to an MCP server. Also use when the conversation involves designing tool schemas, writing tool descriptions, or…
debt-ops-init
Write or refresh a "Tech debt operations" section in the project's AGENTS.md so the team shares one source of truth for debt-ops disciplines. Run ONLY when the user explicitly asks to set up, install, or initialize debt-ops disciplines — never auto-invoke. Idempotent; only the managed section changes, other sections…
debt-ops-metrics
Print a debt-ops health summary from the metrics log, covering registration rate, feedback action rate, ADR creation, and AI-authored share. Use when the user asks for "debt-ops metrics", "debt health", "registry stats", or a tech-debt health summary. Read-only, never writes the log.