Enforce mindful, checklist-driven AI coding discipline via /kasi-* commands and specialist agents for code reviews, security audits, backend fixes, DevOps runbooks, refactors, and knowledge management. Orchestrates multi-model tiers (Opus/Sonnet/Haiku), builds project indexes/graphs, syncs global patterns, and captures replayable session histories in Markdown.
Capture the steps just performed (deploy/migration/hotfix/etc.) as a replayable runbook. Writes to .kasidit/knowledge/runbooks/.
Backend mission — multi-mode (fix/audit/scaffold/design/perf) for API/controller/service/route/DB work
Tier cascade orchestration — Opus plans, Sonnet works, Haiku greps
DevOps mission — deploy flow, infra, CI/CD, env vars, data structure connections, runbooks
Fetch and cache version-matched official documentation
Plan BEFORE code for new features, significant refactors, or changes touching more than 2 files. Returns step-by-step plan, file list, data flow, trade-offs. Does NOT write code.
Multi-focus audit agent — review code quality, security, or performance depending on --focus flag. Replaces and supersedes separate code-reviewer, security-auditor, perf-profiler agents (v0.10). Use --focus=quality|security|perf|all.
Hunt down root cause of a reported bug. Use when test breaks, endpoint returns wrong data, stack trace appears, or user says "why does X fail". Returns minimal-change fix with reproduction.
Deprecated stub (v0.10). Delegated to audit-specialist --focus=quality. Kept for name resolution only; removed in v0.11.
Deep investigation across docs, source, web for library/API/framework questions. Use when task needs version-matched docs, competitor analysis, or codebase survey before deciding. Returns findings + sources, never writes final code.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Stop fighting your AI. Give it discipline instead. Mindful AI coding framework for Claude Code. Works the same on Opus, Sonnet, and Haiku.
Site: kasi web · Wiki: GitHub Wiki · Changelog: CHANGELOG.md
Your AI:
These aren't model bugs. They're discipline gaps — AI lacks a grounded base, so it pattern-matches surface similarity and compounds errors.
The fix is not a bigger model. The fix is scaffolding: checklists, confidence labels, tier-aware rules, runtime verification, mission-driven scope.
| Most prompt frameworks | Kasidit | |
|---|---|---|
| Vague request like "review this" | proceeds and guesses | refuses, demands narrowing |
AI claims [high] confidence | trusts it | runtime hook downgrades it if no Read/Bash call matches |
| Master agent spawns helpers, then edits files itself | nobody notices | hook flags master-orchestrator violation |
| Runs the same on Haiku as on Opus | "Haiku is too small" | tier-aware rules — Haiku gets MORE scaffolding, not less |
| Convention-only rules | model decides to ignore them | 5 runtime-enforced hooks + 16 prompt-level rules |
| Knowledge resets every session | one-shot prompts | Gravity hub persists patterns / runbooks / route-memory across projects |
| One audit = one whole-repo grep | slow + noisy | subgraph extraction — scope to 10 functions, not 1000 |
| Deploy procedure forgotten | re-figure it next time | /kasi-acknowledge captures it as a replayable runbook |
The real edge: discipline turns Haiku into a senior engineer's executor. On Opus it unlocks depth. On Haiku it replaces raw reasoning with checklists, verifier passes, and forced abstention. Same framework, every tier.
/kasi-review · security · fix · ui · backend · graph · struc · devopt · acknowledge · knowledge-list · cascade · init · scaffold · docs · status · promote · pull · sync · search · multi · wiki-sync/kasi mode command — pick intensity: off / router / lite / full / ultraarchitect-planner, audit-specialist (--focus=quality|security|perf), bug-hunter, deep-researcher, legacy-specialist, migration-specialist, refactor-surgeon, test-writer> /kasi-fix the user export endpoint returns null on weekends
[kasidit] kind=bug-fix mode=lite history=12/14 avg_turns=2.1
Mission: user export returns null on weekends
Hypothesis: timezone offset on the cutoff filter — Sat/Sun rolls into Mon
Fix: app/Exports/UserExport.php:42 — use ->whereDate(..., $tz) not ->whereDay
Confidence: [medium] — pattern match, not yet runtime-verified
Verify: curl /api/users/export?on=2026-04-26 → expect 142 rows
# In Claude Code
/plugin marketplace add kasidit-wansudon/kasidit
/plugin install kasidit@kasidit
Or clone and run bash plugins/kasidit/install.sh.
Kasidit also runs on thClaws — the native Rust agent harness:
git clone https://github.com/kasidit-wansudon/kasidit.git
cd kasidit
bash plugins/kasidit/install-thclaws.sh
~85% feature parity (4/5 hooks adapted to thClaws events; full SKILL.md + 21 commands + 8 agents + 15 checklists). See docs/thclaws-setup.md for hook event mapping and limitations.
Already installed? Pull the latest release:
npx claudepluginhub kasidit-wansudon/kasidit --plugin kasiditPersonal Claude Code + Codex dev stack: security hooks, AI-first code conventions, /security-review, /repo-map, /stack-check, portable statusline. Designed to complement other skills-based plugins, not replace them.
AI-powered development workflow automation - Phase-based planning, implementation orchestration, preflight code quality checks with security scanning, ship-it workflow, and development principles generator for CLAUDE.md
Expert code review specialist. Proactively reviews code for quality, security, and maintainability. Use immediately after writing or modifying code.
Turn on Godmode for Claude Code. 126 skills. 7 subagents. Zero configuration.
Code transformation: Dev SDLC orchestrator (code-shipping pipeline), plan, assert, audit, review, test, refactor, debug, for-sure. Hosts engineering agents.
Use this agent when designing APIs, building server-side logic, implementing databases, or architecting scalable backend systems. This agent specializes in creating robust, secure, and performant backend services. Examples:\n\n<example>\nContext: Designing a new API\nuser: "We need an API for our social sharing feature"\nassistant: "I'll design a RESTful API with proper authentication and rate limiting. Let me use the backend-architect agent to create a scalable backend architecture."\n<commentary>\nAPI design requires careful consideration of security, scalability, and maintainability.\n</commentary>\n</example>\n\n<example>\nContext: Database design and optimization\nuser: "Our queries are getting slow as we scale"\nassistant: "Database performance is critical at scale. I'll use the backend-architect agent to optimize queries and implement proper indexing strategies."\n<commentary>\nDatabase optimization requires deep understanding of query patterns and indexing strategies.\n</commentary>\n</example>\n\n<example>\nContext: Implementing authentication system\nuser: "Add OAuth2 login with Google and GitHub"\nassistant: "I'll implement secure OAuth2 authentication. Let me use the backend-architect agent to ensure proper token handling and security measures."\n<commentary>\nAuthentication systems require careful security considerations and proper implementation.\n</commentary>\n</example>