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 neurarelay/relay-action-card --skill neura-relay-coregit clone --depth 1 https://github.com/neurarelay/relay-action-cardWrote 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/neurarelay/relay-action-card/neura-relay-core)<a href="https://agentmods.dev/skills/neurarelay/relay-action-card/neura-relay-core"><img src="https://agentmods.dev/badge/skills/neurarelay/relay-action-card/neura-relay-core/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/neurarelay/relay-action-card/neura-relay-core"><img src="https://agentmods.dev/badge/skills/neurarelay/relay-action-card/neura-relay-core.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 | $0.00058 | $0.00802 |
| Opus 5.5 | $0.00023 | $0.00321 |
| Sonnet 5.5 | $0.00012 | $0.00160 |
| Haiku 4.5 | $0.00006 | $0.00080 |
Grade A, and why
neura-openclaw-core 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 1mo 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 — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Neura Relay Core For OpenClaw-Style Agents
Use this skill before a local autonomous agent takes a consequential action. The goal is one governed checkpoint:
proposed action -> refs-only Action Card -> Relay Decision Receipt -> runtime-owned execution
This is a community example skill pack. It is not an official OpenClaw or ClawHub integration, approval, endorsement, or partnership.
Current Package Truth
- Canonical npm package:
@neurarelay/[email protected] - Canonical ClawHub community package:
@neurarelay/[email protected] - Historical ClawHub community fallback:
@rpelevin/[email protected] - ClawHub history thread:
openclaw/clawhub#2190 - Execution owner: developer or local runtime, not Neura
Prefer the canonical npm package for JavaScript imports. Use the canonical ClawHub community package for OpenClaw installs when ClawHub audits and local policy allow it. The @rpelevin package remains a historical fallback only; neither package creates an official OpenClaw or ClawHub approval, listing, endorsement, integration, or partnership claim.
Workflow
- Classify the action family: message, file, browser, shell, package/publisher, workflow, memory, or data export.
- Build a refs-only Action Card. Include action refs, target refs, authority refs, evidence refs, rule refs, risk category, and requested outcome.
- Reject raw payloads. Do not include message bodies, file contents, form values, raw command strings, package credentials, customer data, secrets, tokens, passwords, or private payloads.
- Route the card through Relay or the preflight adapter. Use dry-run mode when only shaping the card.
- Interpret the Decision Receipt before execution:
proceed: execute only if authority is Registry-backed or the local policy explicitly accepts the receipt route.human_review: stop and collect a named human review ref.revise: update the Action Card with missing evidence, policy, target, or authority refs.stoporblocked: do not execute; record the blocking reason and readiness path.
- Preserve the trace, transaction ref, receipt id, and developer-owned next step.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 1mo ago First seen · 65 lines · 58 tokens per session scan A a6c12585d20f
neura-openclaw-core is a skill published in the GitHub repository neurarelay/relay-action-card (2 stars, last pushed 2mo ago), licensed MIT. It adds 58 tokens to every session and 802 once invoked, about $0.0002 per session on Opus 5.5. 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
markitdown
Convert documents (PDF, Word, Excel, PowerPoint, images) to Markdown text for LLM processing.
markus-hub-connector
Search, download, and publish agents/teams/skills on Markus Hub.
feishu-interaction
Feishu/Lark platform interaction via MCP - documents, tasks, calendar, Bitable, messaging.
chrome-devtools
Chrome DevTools browser automation via MCP - control a live Chrome browser.
sofagent-fde
A Chinese-language deployment and knowledge-engineering process for mapping a company's work and turning its business rules into data and constraints for an AI system. It covers business workflows, possible AI tasks, relationships between concepts, and handover.
adr-create
Use when capturing an architectural decision — file naming, next ADR number, Status / Context / Decision / Consequences, index regen; fires even without saying 'ADR'.