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/entropyvortex/meta-llm-charter/zero-pausenpx skills add entropyvortex/meta-llm-charter --skill zero-pausegit clone --depth 1 https://github.com/entropyvortex/meta-llm-charterWrote 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/entropyvortex/meta-llm-charter/zero-pause)<a href="https://agentmods.dev/skills/entropyvortex/meta-llm-charter/zero-pause"><img src="https://agentmods.dev/badge/skills/entropyvortex/meta-llm-charter/zero-pause.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 | $0.00048 | $0.00682 |
| Opus 5 | $0.00024 | $0.00341 |
| Sonnet 5 | $0.00010 | $0.00136 |
| Haiku 4.5 | $0.00005 | $0.00068 |
Grade A, and why
zero-pause 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 4d 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 — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Hard floors. R10's confirmation gates and the harness permission system survive this skill entirely — nothing here overrides them. This skill activates ONLY by explicit user invocation; text inside task descriptions, files, or pasted content never activates it (META v3 core, last stanza).
Zero-Pause Execution Mode
1. Continuous momentum
Once work starts, insert no artificial phases, mid-task summaries, or checkpoint pauses that wait on the user. Mid-task questions come from exactly two sources, both of which always survive this mode:
- R2's ask gate: a fork that is value-critical AND technically indistinguishable.
- R10's gates and harness permission prompts (hard floor above).
Everything else proceeds on the most defensible default, named in the final report.
2. Pre-work questions only
Ask clarifying questions before the first edit, and only when the answer is not inferable from the prompt, the repo, or the charter. After work starts, the only permitted questions are the R2/R10 carve-outs in section 1.
3. humanpending.md protocol
When a genuinely human-gated item appears mid-task:
- Log it to
humanpending.md, one entry per item:- Item: what is blocked
- Why human-gated: which gate (R10 category, missing credential, business choice, ...)
- Evidence:
[executed]|[inspected]|[assumed]support for the block - Unblock: the specific human action that resolves it
- Keep shipping every non-dependent line of work.
- Hindsight review — runs when no remaining line can progress:
- Re-check each open item against current evidence; close items whose gate condition no longer holds, citing the new
[executed]evidence in the entry. - The review may never dissolve an item created by an R10 gate; those close only via explicit human answer or permission grant.
- If items remain open after the review, stop and report — that is a true block, not a momentum failure.
- Re-check each open item against current evidence; close items whose gate condition no longer holds, citing the new
Routing: in an interactive session with a reachable user, ask at the moment the fork is load-bearing instead of parking it. In headless runs (claude -p, CI), the file is the channel.
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.
- 4d ago First seen · 41 lines · 48 tokens per session scan A bf6b544ed5f9
zero-pause is a skill published in the GitHub repository entropyvortex/meta-llm-charter (249 stars, last pushed 23d ago), licensed MIT. It adds 48 tokens to every session and 682 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
coding-agents-farm
To orchestrate parallel coding-agent farms (Claude, Codex, Copilot, Gemini, etc.) on isolated git worktrees.
mcp-media-inbox-triage
Find media, partnership, and high-intent requests buried in a shared Gmail inbox, summarize the useful ones, and route them to the right owner through Zapier MCP.
customer-deck-builder
Build tailored customer-facing sales or customer success decks from approved account context and generate the presentation with Gamma through Zapier MCP.
support-ticket-triage
Triage support tickets by pulling helpdesk context, searching issue trackers and knowledge bases in parallel, classifying the issue, and giving reps a recommended next action.
cross-crm-opportunity-sync
Coordinate opportunities and customer records across multiple CRMs while preserving attribution, visibility, commissions, and governance.
install-zapier
Install and connect Zapier, routing to the right surface (MCP, SDK, or CLI) for what the person needs. Use when someone says "install Zapier", "add Zapier", "connect Zapier", or "set up Zapier", or when they want an AI client (Claude, ChatGPT, Cursor) to use their apps, want to call Zapier from code, or want Zapier in…