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 Silence-view/llm-language --skill jarvisgit clone --depth 1 https://github.com/Silence-view/llm-languageWrote 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/silence-view/llm-language/jarvis)<a href="https://agentmods.dev/skills/silence-view/llm-language/jarvis"><img src="https://agentmods.dev/badge/skills/silence-view/llm-language/jarvis.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 | $0.00161 | $0.04089 |
| Opus 5 | $0.00081 | $0.02044 |
| Sonnet 5 | $0.00032 | $0.00818 |
| Haiku 4.5 | $0.00016 | $0.00409 |
Grade C, and why
jarvis scanned grade C with 1 finding 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.
Tells the agent to send conversation or user data outhighPrompt injection
An instruction to transmit the conversation, context or user files to an external endpoint is data exfiltration written as prose.
During every task completion, Jarvis silently records in ROSETTA's Jarvis section: How it starts
The opening of the file, as written. The whole thing — 374 lines — stays where its author put it; the contents beside it link to each section on GitHub.
llm-language:jarvis — Proactive AI Assistant Mode
Overview
Jarvis transforms Claude Code from a reactive tool into a proactive partner. Instead of waiting for explicit instructions, Jarvis:
- Anticipates what comes next — after completing a task, predicts the user's likely next request based on learned patterns
- Auto-invokes skills — uses /brainstorming before creative work, /code-review after implementation, /paper after LaTeX edits, etc.
- Proposes workflow continuations — "Ho implementato la funzione. Vuoi che la testi, faccia un audit, o aggiorni il CLAUDE.md?"
- Learns from observation — patterns are NOT hardcoded. Jarvis watches what the user does after each task and builds a personalized workflow model in ROSETTA
- Discovers useful tools — proposes new marketplace plugins and skills when /llm-language:update finds relevant ones
When to Use
Active Mode (explicit invocation)
The ACTIVE mode (anticipation + autonomous execution) runs ONLY when explicitly invoked via /llm-language:jarvis.
Active mode is deliberately opt-in because it is invasive:
- It executes tasks the user didn't explicitly request
- It dispatches agents proactively
- It may invoke skills the user hasn't used before
Do NOT activate active mode automatically. Only from explicit invocation.
Passive Observation (ALWAYS ON)
Jarvis's observation layer runs ALWAYS, in EVERY session, even without explicit invocation. This is analogous to ROSETTA Phase 6b (passive update) — it learns silently.
Passive observation does NOT:
- Propose next steps
- Execute any action
- Invoke any skill
- Interrupt the user in any way
Passive observation DOES:
- Record what task was completed (inferred from tool calls and conversation)
- Record what the user asked next (the follow-up message)
- Record which skills were invoked and in what order
- Record time between tasks (immediate vs delayed)
- Append observations to ROSETTA § Jarvis Patterns silently
This is non-invasive and zero-overhead — it piggybacks on the existing ROSETTA Phase 6b update mechanism. The only difference is that it also logs the task→next_action sequence, not just user preferences.
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 · 374 lines · 161 tokens per session scan C aadcb56e8b30
jarvis is a skill published in the GitHub repository Silence-view/llm-language (2 stars, last pushed 2mo ago), licensed MIT. It adds 161 tokens to every session and 4,089 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it C with 1 finding (tells the agent to send conversation or user data out). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
cco-budget
Configure token budget limits, auto-compact settings, and view current budget status (model-aware — Claude 5 lineup, Opus 5 default fallback, full 1M context at standard price).
cco-clean
Clean up old tracking data and reset statistics.
nightly
Canonical seven-leg nightly close-parity protocol. Eleven independently evidenced checkpoints cover synthesis, reconciliation, context hygiene, sweeps, heat, digest, runtime health, cognitive state, ledger capture/review, and vault sync.
Promote
Take a slice of the current conversation and promote it to a durable vault artifact — concept note, project update, person profile, or doctrine entry. Pairs with /close. /close writes the daily note (transient outcome record); /promote elevates specific insights to permanent vault content the framework can refer back…
Stuck
Walk the 6-rung escalation ladder before declaring stuck or asking the principal. Most "stuck" is actually "haven't tried the next approach." Stopping is the last resort, not the first response.
open
Boot the session, load context from the vault, and surface what matters.