eyalgo/kensaiapp
Built as an agentic software development system on Claude Code, from requirements to merged code. Shared openly, as a working experiment.
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
GUARDRAILS
3/5
Blocks destructive commands · Pre-tool screening hook · No YOLO mode · details
Quality 18/100 Guardrails v2
Copy this rig
# review before running: this installs third-party code
$ npx degit eyalgo/kensaiapp/.claude ./rig-kensaiapp # inspect, then merge into .claude/ MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ npx degit eyalgo/kensaiapp/skills/architect/.claude/skills/development-design .claude/skills/development-design $ npx degit eyalgo/kensaiapp/skills/cdo/.claude/skills/experience-visual .claude/skills/experience-visual $ npx degit eyalgo/kensaiapp/skills/cdo/.claude/skills/product-design .claude/skills/product-design $ npx degit eyalgo/kensaiapp/skills/cdo/.claude/skills/product-uxd .claude/skills/product-uxd $ npx degit eyalgo/kensaiapp/skills/companion/.claude/skills/companion-self-answer .claude/skills/companion-self-answer $ npx degit eyalgo/kensaiapp/skills/content-editor/.claude/skills/changelog-line .claude/skills/changelog-line $ npx degit eyalgo/kensaiapp/skills/content-editor/.claude/skills/development-content .claude/skills/development-content $ npx degit eyalgo/kensaiapp/skills/content-manager/.claude/skills/experience-voice .claude/skills/experience-voice $ npx degit eyalgo/kensaiapp/skills/content-manager/.claude/skills/kensaiapp-writing-bar .claude/skills/kensaiapp-writing-bar $ npx degit eyalgo/kensaiapp/skills/content-manager/.claude/skills/product-content .claude/skills/product-content $ npx degit eyalgo/kensaiapp/skills/cpo/.claude/skills/first-what-left .claude/skills/first-what-left $ npx degit eyalgo/kensaiapp/skills/cpo/.claude/skills/product-prd .claude/skills/product-prd $ npx degit eyalgo/kensaiapp/skills/cpo/.claude/skills/project-finish-line .claude/skills/project-finish-line $ npx degit eyalgo/kensaiapp/skills/cto/.claude/skills/product-sdd .claude/skills/product-sdd $ npx degit eyalgo/kensaiapp/skills/cto/.claude/skills/product-srs .claude/skills/product-srs $ npx degit eyalgo/kensaiapp/skills/developer/.claude/skills/debugging-and-error-recovery .claude/skills/debugging-and-error-recovery $ npx degit eyalgo/kensaiapp/skills/developer/.claude/skills/development-implementation .claude/skills/development-implementation $ npx degit eyalgo/kensaiapp/skills/developer/.claude/skills/refactor-for-clarity .claude/skills/refactor-for-clarity $ npx degit eyalgo/kensaiapp/skills/developer/.claude/skills/refactor-for-optimize .claude/skills/refactor-for-optimize $ npx degit eyalgo/kensaiapp/skills/developer/.claude/skills/refactor-for-reuse .claude/skills/refactor-for-reuse $ npx degit eyalgo/kensaiapp/skills/developer/.claude/skills/source-driven-development .claude/skills/source-driven-development $ npx degit eyalgo/kensaiapp/skills/developer/.claude/skills/test-driven-development .claude/skills/test-driven-development $ npx degit eyalgo/kensaiapp/skills/developer/.claude/skills/verify-before-claim .claude/skills/verify-before-claim $ npx degit eyalgo/kensaiapp/skills/experimental/.claude/skills/kaizen-audit .claude/skills/kaizen-audit $ npx degit eyalgo/kensaiapp/skills/experimental/.claude/skills/skill-evals .claude/skills/skill-evals $ npx degit eyalgo/kensaiapp/skills/interactive-product/.claude/skills/kensaiapp-michi .claude/skills/kensaiapp-michi $ npx degit eyalgo/kensaiapp/skills/interactive-product/.claude/skills/product-card .claude/skills/product-card $ npx degit eyalgo/kensaiapp/skills/interactive-product/.claude/skills/project-setup .claude/skills/project-setup $ npx degit eyalgo/kensaiapp/skills/interactive-product/.claude/skills/project-stamp .claude/skills/project-stamp $ npx degit eyalgo/kensaiapp/skills/interactive/.claude/skills/agent-invoke .claude/skills/agent-invoke $ npx degit eyalgo/kensaiapp/skills/interactive/.claude/skills/companion-answer .claude/skills/companion-answer $ npx degit eyalgo/kensaiapp/skills/interactive/.claude/skills/session-close .claude/skills/session-close $ npx degit eyalgo/kensaiapp/skills/interactive/.claude/skills/skill-creator .claude/skills/skill-creator $ npx degit eyalgo/kensaiapp/skills/interactive/.claude/skills/skill-evals-creator .claude/skills/skill-evals-creator $ npx degit eyalgo/kensaiapp/skills/interactive/.claude/skills/skill-evals-trigger .claude/skills/skill-evals-trigger $ npx degit eyalgo/kensaiapp/skills/kensaiapp-auditor/.claude/skills/development-audit .claude/skills/development-audit $ npx degit eyalgo/kensaiapp/skills/pm/.claude/skills/nemawashi-unpick .claude/skills/nemawashi-unpick $ npx degit eyalgo/kensaiapp/skills/pm/.claude/skills/seed-pick .claude/skills/seed-pick $ npx degit eyalgo/kensaiapp/skills/pm/.claude/skills/shukuba-open .claude/skills/shukuba-open $ npx degit eyalgo/kensaiapp/skills/pm/.claude/skills/shukuba-park .claude/skills/shukuba-park
$ curl -fsSL --create-dirs -o .claude/agents/architect.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/architect/.claude/agents/architect.md $ curl -fsSL --create-dirs -o .claude/agents/cdo.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/cdo/.claude/agents/cdo.md $ curl -fsSL --create-dirs -o .claude/agents/content-editor.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/content-editor/.claude/agents/content-editor.md $ curl -fsSL --create-dirs -o .claude/agents/content-manager.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/content-manager/.claude/agents/content-manager.md $ curl -fsSL --create-dirs -o .claude/agents/cpo.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/cpo/.claude/agents/cpo.md $ curl -fsSL --create-dirs -o .claude/agents/cto.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/cto/.claude/agents/cto.md $ curl -fsSL --create-dirs -o .claude/agents/developer.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/developer/.claude/agents/developer.md $ curl -fsSL --create-dirs -o .claude/agents/delegate.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/interactive/.claude/agents/delegate.md $ curl -fsSL --create-dirs -o .claude/agents/kaizen-fix-author.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/kaizen-fix-author/.claude/agents/kaizen-fix-author.md $ curl -fsSL --create-dirs -o .claude/agents/kaizen-fixer.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/kaizen-fixer/.claude/agents/kaizen-fixer.md $ curl -fsSL --create-dirs -o .claude/agents/kensaiapp-auditor.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/kensaiapp-auditor/.claude/agents/kensaiapp-auditor.md $ curl -fsSL --create-dirs -o .claude/agents/pm.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/pm/.claude/agents/pm.md $ curl -fsSL --create-dirs -o .claude/agents/principal-architect.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/principal-architect/.claude/agents/principal-architect.md $ curl -fsSL --create-dirs -o .claude/agents/principal-pm.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/principal-pm/.claude/agents/principal-pm.md $ curl -fsSL --create-dirs -o .claude/agents/product-owner.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/product-owner/.claude/agents/product-owner.md $ curl -fsSL --create-dirs -o .claude/agents/qa.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/qa/.claude/agents/qa.md $ curl -fsSL --create-dirs -o .claude/agents/scrum-master.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/scrum-master/.claude/agents/scrum-master.md $ curl -fsSL --create-dirs -o .claude/agents/senior-engineer.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/senior-engineer/.claude/agents/senior-engineer.md $ curl -fsSL --create-dirs -o .claude/agents/tech-lead.md https://raw.githubusercontent.com/eyalgo/kensaiapp/main/skills/tech-lead/.claude/agents/tech-lead.md
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"hooks": {
"PreToolUse": [
{
"matcher": "Bash",
"hooks": [
{
"type": "command",
"command": "\"${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh\" \"${CLAUDE_PROJECT_DIR}/.claude/hooks/development-bash-approve.py\""
}
]
},
{
"matcher": "Edit|Write",
"hooks": [
{
"type": "command",
"command": "\"${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh\" \"${CLAUDE_PROJECT_DIR}/.claude/hooks/written_forward_guard.py\""
}
]
},
{
"matcher": "Agent",
"hooks": [
{
"type": "command",
"command": "\"${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh\" \"${CLAUDE_PROJECT_DIR}/.claude/hooks/brief_anchor_guard.py\""
}
]
},
{
"matcher": "*",
"hooks": [
{
"type": "command",
"command": "\"${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh\" \"${CLAUDE_PROJECT_DIR}/.claude/hooks/read_first_guard.py\""
}
]
}
]
}
} Skills (59)
development-designexperience-visualproduct-designproduct-uxdcompanion-self-answerchangelog-linedevelopment-contentexperience-voicekensaiapp-writing-barproduct-contentfirst-what-leftproduct-prdproject-finish-lineproduct-sddproduct-srsdebugging-and-error-recoverydevelopment-implementationrefactor-for-clarityrefactor-for-optimizerefactor-for-reusesource-driven-developmenttest-driven-developmentverify-before-claimkaizen-auditskill-evalskensaiapp-michiproduct-cardproject-setupproject-stampagent-invokecompanion-answersession-closeskill-creatorskill-evals-creatorskill-evals-triggerdevelopment-auditnemawashi-unpickseed-pickshukuba-openshukuba-parkshukuba-refinestations-from-seedtobiishi-droptobiishi-opentobiishi-sealproduct-architecturenemawashi-dropnemawashi-refineproduct-fdrseeds-from-ideaseeds-from-product-docsdevelopment-specsnemawashi-retireshukuba-retiredependency-repairdevelopment-code-integritydevelopment-securityad-hoc-plandevelopment-plan
Subagents (19)
| architect | Translate a spec into system design — module boundaries, integration points, build sequence, change_design.md. Use for architecture/design work, not implementation. |
| cdo | Own experience definition at the executive tier — turn the CPO's PRD into the UX design specification (UXD.md) and the design language (DESIGN.md) that the SDD and implementing agents consume, upstrea |
| content-editor | Write the text one change publishes, to the standard the Content Manager keeps and to the product's CONTENT.md and the instance's writing bar. Distinct from the Content Manager, who owns the standing |
| content-manager | Own the product's words — whether content is needed, where it belongs, and how it is worded — and author CONTENT.md, the verbal language that is the peer of the CDO's DESIGN.md. Distinct from the CDO, |
| cpo | Own product intent at the executive tier and author the PRD — turn a decided product idea into the product source of truth, upstream of the change pipeline. Distinct from the delivery-tier Product Own |
| cto | Own technical definition at the executive tier — turn the CPO's PRD into the SRS (and co-lead the SDD), upstream of the change pipeline. Distinct from the delivery-tier Architect who designs each indi |
| developer | Implement specs with TDD and clean code — write features, fix bugs, write tests, commit. The spec defines what to build; you own how well it is built. |
| delegate | The agent an owner's one-off delegated work runs as — research, an investigation, a design or plan, a bounded change. Reach for it when the work is multi-step and needs to WRITE something (a document, |
| kaizen-fix-author | Decides what a skill's fix is and writes its exact replacement text — the deciding half of the fix, paired with the kaizen-fixer that applies it and the kensaiapp-auditor that only sees. Its task prom |
| kaizen-fixer | Applies a decided fix to a skill's stack and commits the round — the acting half, paired with the kensaiapp-auditor reader that only sees and the kaizen-fix-author that decides. Subtracts or trades, n |
| kensaiapp-auditor | Fresh, role-neutral eyes over an artifact it did not produce — enumerates the defects in KensaiApp output and emits findings, never fixes. Two jobs, set by the task prompt. Flow-audit (launched by the |
| pm | Own one project's traverse at the delivery tier — pick the seed that is worked now, cut it into the stations it will take, and open each station and each change unit. Distinct from the executive-tier |
| principal-architect | Own the system's structure and its constraints at the executive tier — record the architecturally significant decisions as an append-only log, and bind what a later change may not break, downstream of |
| principal-pm | Own the program's execution shape at the executive tier — author a product's FDR once its SDD settles, and cut into the journey's seeds and their station checklists, upstream of the PM's loop. Two cut |
| product-owner | Own the specification of one change at the delivery tier — take the PM's change request and turn it into change_spec.md through development-specs, owning scope, behaviour, business rules and acceptanc |
| qa | Independently verify committed work with fresh eyes — run acceptance checks, exercise tests, write verification artifacts. Use for verification, not authoring. |
| scrum-master | The delivery-tier process seat, distinct from the PM, who owns a project's traverse and the moves the journey waits on. This seat closes a reached station with shukuba-retire and retires a walked seed |
| senior-engineer | The delivery-tier engineer — proves finished code does what its intent says (the code-integrity step), reviews a change's branch for security (the security skill), and merges a reviewed ad-hoc branch. |
| tech-lead | Turn a change design into an executable tl_instruction.md and review code before runs that spend budget or exercise the system end-to-end. Use for build planning and pre-run review. |
Hooks (15)
| event | matcher | runs |
|---|---|---|
| SessionStart | * | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/resolve_core.py" |
| SessionStart | * | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/task_map_salience.py" |
| SessionStart | * | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/manifest_salience.py" |
| SessionStart | * | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/standing_rules_salience.py" |
| SubagentStart | * | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/resolve_core.py" |
| SubagentStart | * | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/task_map_salience.py" |
| SubagentStart | * | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/manifest_salience.py" |
| SubagentStart | * | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/standing_rules_salience.py" |
| SubagentStop | delegate | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/subagent_stop_notify.py" |
| PreToolUse | Bash | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/development-bash-approve.py" |
| PreToolUse | Edit|Write | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/written_forward_guard.py" |
| PreToolUse | Agent | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/brief_anchor_guard.py" |
| PreToolUse | * | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/read_first_guard.py" |
| PostToolUse | Bash | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/context_cap.py" |
| UserPromptSubmit | * | "${CLAUDE_PROJECT_DIR}/core/tools/locked_python.sh" "${CLAUDE_PROJECT_DIR}/.claude/hooks/permission_mode_reminder.py" |
Similar rigs
yeyo11/agentry
Run Claude Code as a service: REST API, web UI and multi-agent orchestration in one container
Pragmatist 5.2k tok · · Q18
ThinhHoang1/claude-code-agent-team
An AI dev team for Claude Code agent teams: 8 role agents, a HANDOFF protocol and 3 hook quality gates — with the TeammateIdle infinite-loop bug fixed and regression-tested.
Orchestrator 1.6k tok · · Q15
yeyo11/claudio
Run Claude Code as a service: a REST API, a web UI, a Linux desktop app and parallel agents on a project's board, all through the Claude Code CLI.
Orchestrator 6.0k tok · · Q18
danielfarnose/claude-code-dev-loop
A multi-agent development loop for Claude Code: five specialized subagents plan, implement, verify and merge — with a QA gate that cannot edit the code it judges.
Orchestrator 1.2k tok · · Q0