asklokeshasklokeshCommunityยท3 Agent Skills Included

loki-mode

Autonomous spec-to-product builder with verified completion evidence

Turns any spec (PRD, GitHub issue, OpenAPI doc, or one-line brief) into a tested, deployable product through autonomous Reason-Act-Reflect-Verify cycles. Eliminates false 'done' claims by blocking completion on empty diffs, failing tests, and leaked secrets, then issuing a re-checkable Evidence Receipt. Runs on Claude Code, Codex, Cline, or Aider with 8 quality gates, blind code review, and budget controls.
npx skills add asklokesh/loki-mode --all -g -y
Available:

Instructs the AI agent on how to run Loki Mode's autonomous build loop, including provider selection, model tiers, quality gates, memory modules, and skill routing rules.

All Skills in This Repository (3)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install Loki Mode?โ–ผ

Run `npx skills add asklokesh/loki-mode --all -g -y` in your terminal to install all skills in this suite globally.

What does Loki Mode actually build?โ–ผ

It takes any spec such as a PRD, GitHub issue, or OpenAPI document and autonomously builds, tests, and deploys the product with minimal human input.

How does Loki Mode verify work is really done?โ–ผ

It blocks completion on empty diffs, failing tests, or leaked secrets, and writes an Evidence Receipt you can independently re-verify with `loki proof verify`.

Which AI providers does Loki Mode support?โ–ผ

It runs on Claude Code, OpenAI Codex CLI, Cline, and Aider, with full features on Claude and graceful degraded modes on the others.

Can I use Loki Mode on an existing codebase?โ–ผ

Yes. Run `loki modernize heal ./your-repo --assess` for a read-only assessment that changes nothing before you commit to any modifications.

Related Repositories in Software Engineering

View All in Software Engineeringโ†’