# Teamwork Goal

> Use when the user asks to keep going, run until it passes, fix until green, iterate until done, converge on a verifiable target, or work within an explicit budget.

- **Type:** Skill
- **Install:** `agentstack add skill-jinplu-teamwork-teamwork-goal`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [JinPLu](https://agentstack.voostack.com/s/jinplu)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [JinPLu](https://github.com/JinPLu)
- **Source:** https://github.com/JinPLu/Teamwork/tree/main/skills/teamwork-goal

## Install

```sh
agentstack add skill-jinplu-teamwork-teamwork-goal
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Teamwork Goal

## Outcome

Iterate until a verifiable target passes, the authorized budget ends, or a hard
stop is reached.

## Enter When

Use for explicit persistence such as “keep going,” “until green,” or a stated
budget. Ordinary one-shot work stays in its narrower stage. If objective,
verification, scope, Goal Invariants, or stop rules are unclear, propose them
and obtain approval before creating goal state.

## Do And Boundaries

Use Codex native goal state when available; otherwise keep a rolling report in
`docs/teamwork/reports/`. Record the target, protected boundaries, Goal
Invariants, verification, and the budget supplied by the user/runtime or
accepted plan. Do not invent a fixed iteration budget.

Before retrying, run Replay Preflight: identify prior attempts, current evidence,
do-not-repeat constraints, and the strategy delta.
Route broad unknowns to research, reproducible unknown-cause failures to debug,
scope decisions to plan, changes to execute, and acceptance to review. After
each verify/review cycle, append the attempt result; failures and no-progress
cycles also record why and what must change. Accept only when focused
verification and the required review pass.

## Done When

Mark complete only after the success signal passes. Otherwise return the
current attempt evidence, remaining budget if defined, next strategy, or the
specific stop reason.

## Escalate

Stop on repeated no-progress without an evidence-backed strategy delta,
destructive risk, auth/resource failure, protected-boundary conflict, or
unresolved user intent. Ask before expanding authority or scope.

## Conditional Protocols

Use `goal-iteration.md` for proposal, replay, attempt, and rolling-report
formats; use each narrower stage protocol when that stage runs. Paths are under
`skills/using-teamwork/references/`.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [JinPLu](https://github.com/JinPLu)
- **Source:** [JinPLu/Teamwork](https://github.com/JinPLu/Teamwork)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-jinplu-teamwork-teamwork-goal
- Seller: https://agentstack.voostack.com/s/jinplu
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.