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.
git clone --depth 1 https://github.com/gaia-research/gaia-skill-heavenWrote 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/commands/gaia-research/gaia-skill-heaven/heaven)<a href="https://agentmods.dev/commands/gaia-research/gaia-skill-heaven/heaven"><img src="https://agentmods.dev/badge/commands/gaia-research/gaia-skill-heaven/heaven/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/commands/gaia-research/gaia-skill-heaven/heaven"><img src="https://agentmods.dev/badge/commands/gaia-research/gaia-skill-heaven/heaven.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 | $0.00011 | $0.00337 |
| Opus 5.5 | $0.00004 | $0.00135 |
| Sonnet 5.5 | $0.00002 | $0.00067 |
| Haiku 4.5 | $0.00001 | $0.00034 |
Grade A, and why
heaven scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Skill Heaven console (preview)
Default output must be the concise bundled summary (at most eight lines). Never add raw JSON, lengthy capability notes, fallback recipes or diagnostic prose. Only an explicit /heaven inspect [status|lens|session|scope|flow|trust] request adds --details and the chosen --surface to the same report invocation. Never run --details as the default. Printed commands are never executed. The host owns dismissal and tool-trace presentation; do not install a persistent statusline.
This command is model-mediated, not a native pane or persistent HUD.
Run node "${GROK_PLUGIN_ROOT}/scripts/heaven.mjs" --host grok and print stdout verbatim. A person may supply --session-root; use only the exact Core sessionRoot returned by the summon tool in THIS conversation. Never look for a newest directory, scan unrelated session logs or infer a session from the working directory. Without a binding the report says unknown, not zero. No raw transcript export, shared config edits or hidden submission. Printed commands are suggestions only.
Only on explicit /heaven lens <need>: call the existing Core skill-summon tool once with query , surface any and preview:true. Never start another summon server or call without preview. Run the report using the exact sessionRoot from that result, if present. The human types a printed /summon suggestion; you never submit it or read a candidate body automatically. Disclose that this preview and report are model-mediated. If Core is missing, say unavailable.
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.
- 8d ago First seen · 10 lines · 11 tokens per session scan A 3bd26bc4cb8b
heaven is a command published in the GitHub repository gaia-research/gaia-skill-heaven (4 stars, last pushed 6d ago), licensed MIT. It adds 11 tokens to every session and 337 once invoked, about $0.0000 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 2026-10-10.
Other commands, from other repositories
mesh-watch
Manually wait for one responder inbox message.
mesh
Invoke the Agent Mesh CLI.
psql-query
Run ad-hoc PostgreSQL analytics queries against dev/test database.
test-feature
Test a React Native feature on the running simulator/emulator. Verifies UI, user flows, and internal state. Generates a persistent Maestro test file.
memory-forget
Command "memory-forget" from waittim/MemoryCustodian, covering memory-forget and review the preview, then apply its plan id.
spec
Generate technical specifications from design artifacts.