jarvis-goal

jarvis-goal is a skill for Claude Code from Ramsbaby/jarvis. It costs 67 tokens per session (662 once invoked), scanned C, original, MIT.

A goal-driven workflow that turns a stated completion condition into a checkable test and keeps working until the test passes. It pauses for explicit approval before actions that cannot be easily undone.

In plain words
What is it for?
Use it for measurable outcomes such as passing a verification checklist, completing data indexing with an integrity check, or maintaining a target success rate. It can pause before publishing, sending messages, moving money, exposing secrets, or deleting data.
Why use it?
It prevents an agent from declaring success based on assumptions and limits risky autonomous operations.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: mentions Claude Code.

Part of the jarvis-goal plugin — 1 skill shipped together

Good fit Use it for measurable outcomes such as passing a verification checklist, completing data indexing with an integrity check, or maintaining a target success rate. It can pause before publishing, sending messages, moving money, exposing secrets, or deleting data.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/ramsbaby/jarvis/goal
Install

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.

Any agent
npx skills add Ramsbaby/jarvis --skill goal
Clone the repo
git clone --depth 1 https://github.com/Ramsbaby/jarvis

Made for: Claude Code.

Or install jarvis-goal, the plugin that ships this one along with the rest of its 1 skill.

Wrote 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.

agentmods badge for jarvis-goal

README.md
[![agentmods](https://agentmods.dev/badge/skills/ramsbaby/jarvis/goal/github.svg)](https://agentmods.dev/skills/ramsbaby/jarvis/goal)
Your own site
<a href="https://agentmods.dev/skills/ramsbaby/jarvis/goal"><img src="https://agentmods.dev/badge/skills/ramsbaby/jarvis/goal/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.

agentmods 80×15 button for jarvis-goal

Your own site · 80×15
<a href="https://agentmods.dev/skills/ramsbaby/jarvis/goal"><img src="https://agentmods.dev/badge/skills/ramsbaby/jarvis/goal.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 67 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 662 The whole file, excluding the scripts and references it only reads on demand.
Security scan C 1 finding. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00067 $0.00662
Opus 5 $0.00034 $0.00331
Sonnet 5 $0.00013 $0.00132
Haiku 4.5 $0.00007 $0.00066

Measured 9d ago against content hash 4072e649a964, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade C, and why

jarvis-goal 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 9d 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.

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

- `rm -rf`, `DROP TABLE`, `TRUNCATE`
plugins/jarvis-goal/skills/goal/SKILL.md · 63 lines

How it starts

The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/goal — Goal-Driven Autonomous Execution

Set a measurable completion condition once; the agent self-verifies every turn until the condition is met, then declares done with evidence.

How it works (4 stages)

  1. Parse completion condition → normalize to a verifiable predicate
  2. Per-turn self-check → after each action, ask "is the condition met?" using actual tool output, not assumption
  3. Auto-continuation → if not met, continue without waiting for Stop hook
  4. Stop on success → declare completion citing real command output as evidence

Example inputs:

  • /goal all crons sustained ≥ 95% success over 7 days
  • /goal /verify 7-gate all PASS
  • /goal rag indexing complete + integrity check passes

⚠️ Irreversibility Guard (auto-pause triggers)

Autonomous execution pauses and requests explicit user approval when detecting:

  • git push, git push --force
  • Repo visibility changes, deletion, archival (gh repo edit --visibility, gh repo delete, gh repo archive)
  • External message dispatch (Slack, email, Discord public channels, push notifications)
  • Payment, financial transfer, asset movement, trade execution
  • Secret or PII exposure to external systems
  • Irreversible system changes (production DB schema, mass data deletion, cron mass disable)
  • rm -rf, DROP TABLE, TRUNCATE

Pause format:

"Irreversible operation detected: [name]. Scope: [blast radius]. Approve to proceed?"

Differentiation vs. Anthropic /goal

  • Built-in irreversibility guard with explicit pause behavior (the official command has no such guard)
  • Optional ledger integration for token accounting per goal
  • Composable with /verify and /doctor for complex condition predicates

Self-check (before declaring completion — BLOCKING)

  1. Did I confirm the evidence via actual tool output, not assumption?
  2. Have I triggered any irreversibility guard pattern unintentionally?
  3. Did I avoid declaring completion based on partial metrics (e.g., only load average without memory/disk)?
  4. Have I considered at least three alternative hypotheses for any failure I diagnosed?

Read the full file on GitHub · 63 lines

Changes

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.

  1. 9d ago First seen · 63 lines · 67 tokens per session scan C 4072e649a964

Subscribe to this mod's changes

jarvis-goal is a skill published in the GitHub repository Ramsbaby/jarvis (16 stars, last pushed 17d ago), licensed MIT. It adds 67 tokens to every session and 662 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

cognition-authority

How to hold the controls of the Cognition Discord server — you are the first actor there, not an assistant waiting for instructions. Read this before any other Cognition skill and before the first action of any session touching guild 1535979812860993617. It settles who decides, how to read a request as intent rather…

justpainful/Cognition · 0 tokens

build-system

A skill for building interactive systems in the Cognition Discord server by combining Registry building blocks instead of writing code. These systems can include tickets, applications, voting, verification, and menus for choosing roles.

justpainful/Cognition · 0 tokens

registry-authoring

A reference for the Cognition Registry, including its action types, parameters, conditions, template variables, and known configuration mistakes.

justpainful/Cognition · 0 tokens

danger

A safety protocol for irreversible operations on a Cognition server, such as deleting a channel, category, or role. It requires taking a snapshot, previewing the change, getting human approval, and then applying it.

justpainful/Cognition · 0 tokens

react-to-events

A skill for making a Cognition Discord server respond automatically to events such as members joining, keywords being posted, reactions, or channels being deleted.

justpainful/Cognition · 0 tokens

run-session

A procedure for running experiments on a Cognition Discord server from a temporary test stage to a permanent or archived result.

justpainful/Cognition · 0 tokens