Kiwipow Agent
A VS Code coding agent that plans a feature before it reads your code, and calls nothing done until a test proves each rule.
Early days. Kiwipow Agent is young and you will hit bugs. Things are changing frequently, use at own risk! Report them; they usually get fixed fast.
Why it's different
|
What it does |
Why you want it |
| Feature planning, blind to the code |
The planner reads your docs and approved specs, never the code (enforced at the tool call), and writes a spec of named rules. |
A planner that reads the code inherits its bugs as requirements; this one plans from intent. |
| You rule on disagreements |
The spec is mapped against the code; each conflict becomes a decision card with both sides and proposed rewordings. |
Nothing is silently absorbed. You decide whether the spec or the code is wrong. |
| Done means proven |
Every task names the test that proves each rule it delivers; the plan view shows each rule's task and test, or the gap. |
No "done" without evidence you can click through. |
| Targeted verification |
When all tasks are tested, your test commands run over just the projects the feature touched; a failure goes back to the implementer. |
Only the suites that matter run, and red never reaches you as finished. |
| Scripts instead of turns |
The model writes one JavaScript program that reads, greps, runs commands and edits across many files, in a sandbox that reaches nothing but those gated functions. Edits are staged and shown as one diff to approve. |
One turn instead of dozens: faster, and only the result enters the context, not every file along the way. |
| Cleanup after green |
Functions, types and files that grew past your limits are split once tests pass. |
The feature lands without leaving a mess. |
| Specs compound |
An approved spec is read by the next planner like a doc. |
What one feature settled reaches the next without restating it. |
| Docs evaluation |
Reads your docs the way the planner does and says where their arrangement costs a plan, then fixes what you pick. |
Better docs, better plans. |
Any model, per step
- Claude through the Claude Agent SDK.
- Any OpenAI-compatible endpoint on Kiwipow Agent's own tool loop.
- A profile picks the model per step: the strongest reasoner for planning and mapping, a fast cheap one for implementation.
- Your
.mcp.json servers (the workspace's and ~/.mcp.json, with Claude Code's user-wide list moved there on first run), CLAUDE.md/AGENTS.md, skills and permission rules work the same on both.
Also
- Chat sessions for everyday work, with the full tool set.
- Plan sessions for work the code shapes, such as a UI on its framework: intent is agreed before the code is read, then the plan is made against the code and built in a chat.
- Prompts that don't nag. Read-only tools and commands, your
package.json scripts and your test commands run without asking; other shell calls are prompted command by command.
- Large JSON without reading it whole: query a file by expression and get back only the rows you asked for.
Get started
Install Kiwipow Agent from the Marketplace. Windows only (x64, arm64).
Claude sessions need one of:
- an Anthropic API key, added to the Claude provider on the settings page (gear icon), or
- a Claude Code login already on this machine, used when no key is set.
Open Kiwipow Agent in the activity bar, press +, pick Feature planning and describe the feature. For a non-Claude model, add a provider on the settings page.
More
| |
