bkit-codex
AGENTS.md69 starsChanged 8 months agoApache-2.0View on GitHub
This project uses bkit with automatic level detection. Call bkitdetectlevel at session start to determine the current level. Starter (beginners, static websites): - Use simple HTML/CSS/JS or Next.js App Router - Skip API and database phases - Pipeline phases: 1 -> 2 -> 3 -> 6 -> 9 - Use $starter skill for beginner guidance Dynamic (fullstack with BaaS): - Use bkend.ai for backend services - Follow phases: 1 -> 2 -> 3 -> 4 -> 5 -> 6 ->…
We split this AGENTS.md into the rules it gives an agent
15 rules. About 0 are common, 7 are distinctive, and 8 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
Distinctive · estimate
This author's own choices. Worth a look.
8
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 6
Call bkit_detect_level at session start to determine the current level.
- Very specific: line 11
Use simple HTML/CSS/JS or Next.js App Router
- Very specific: line 14
Use $starter skill for beginner guidance
- Distinctive: line 17
Use bkend.ai for backend services
- Distinctive: line 18
Follow phases: 1 -> 2 -> 3 -> 4 -> 5 -> 6 -> 7 -> 9 (phase 8 optional)
- Very specific: line 19
Use $dynamic skill for fullstack guidance
- Very specific: line 23
Use $enterprise skill for MSA guidance
- Very specific: line 27
ALWAYS check docs/.pdca-status.json for current feature status.
- Very specific: line 28
Use bkit_get_status MCP tool for parsed status with recommendations.
- Distinctive: line 44
ALWAYS include at the end of each response:
- Distinctive: line 47
Use simple terms, avoid jargon.
- Distinctive: line 63
When working on complex features:
- Distinctive: line 68
Do: Developer + Frontend/Backend perspective
- Distinctive: line 69
Check: QA + Code Reviewer perspective
- Very specific: line 71
Use bkit_pdca_next to transition between phases
≈ “Common” is an estimate: at least 1 in 5 of the 36,368 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
bkit Project Configuration
Project Level
This project uses bkit with automatic level detection.
Call bkit_detect_level at session start to determine the current level.
Level-Specific Guidance
Starter (beginners, static websites):
- Use simple HTML/CSS/JS or Next.js App Router
- Skip API and database phases
- Pipeline phases: 1 -> 2 -> 3 -> 6 -> 9
- Use
$starterskill for beginner guidance
Dynamic (fullstack with BaaS):
- Use bkend.ai for backend services
- Follow phases: 1 -> 2 -> 3 -> 4 -> 5 -> 6 -> 7 -> 9 (phase 8 optional)
- Use
$dynamicskill for fullstack guidance
Enterprise (microservices, K8s):
- All 9 phases required
- Use
$enterpriseskill for MSA guidance
PDCA Status
ALWAYS check docs/.pdca-status.json for current feature status.
Use bkit_get_status MCP tool for parsed status with recommendations.
Key Skills
| Skill | Purpose |
$pdca | Unified PDCA workflow (plan, design, do, analyze, iterate, report) |
$plan-plus | Brainstorming-enhanced planning (6 phases, HARD GATE) |
$starter / $dynamic / $enterprise | Level-specific guidance |
$development-pipeline | 9-phase pipeline overview |
$code-review | Code quality analysis with static analysis patterns |
$bkit-templates | PDCA document template selection |
Response Format (MANDATORY)
Starter Level (bkit-learning style)
ALWAYS include at the end of each response:
- Learning Points: 3-5 key concepts the user should learn
- Next Learning Step: What to study or practice next
- Use simple terms, avoid jargon. Use "Did you know?" callouts.
Dynamic Level (bkit-pdca-guide style)
ALWAYS include at the end of each response:
- PDCA Status Badge:
[Feature: X | Phase: Y | Progress: Z%] - Checklist: What's done and what remains
- Next Step: Specific action with command/tool suggestion
Enterprise Level (bkit-enterprise style)
ALWAYS include at the end of each response:
- Tradeoff Analysis: Pros/Cons of the approach taken
- Cost Impact: Development time, infrastructure cost, maintenance burden
- Deployment Considerations: Environment-specific notes
Team Workflow (Single Agent Mode)
When working on complex features:
- Break the task into PDCA phases (Plan -> Design -> Do -> Check -> Report)
- For each phase, apply the relevant specialist perspective:
- Plan: Product Manager + CTO perspective
- Design: Architect + Security perspective
- Do: Developer + Frontend/Backend perspective
- Check: QA + Code Reviewer perspective
- Report: Documentation perspective
- Use
bkit_pdca_nextto transition between phases - Quality gates: Each phase must be documented before proceeding
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 28,693 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository27 other files it gives its agents.
- bkend-authSkill
- bkend-cookbookSkill
- bkend-dataSkill
- bkend-quickstartSkill
- bkend-storageSkill
- bkit-rulesSkill
- bkit-templatesSkill
- code-reviewSkill
- codex-learningSkill
- desktop-appSkill
- development-pipelineSkill
- dynamicSkill
- enterpriseSkill
- mobile-appSkill
- pdcaSkill
- phase-1-schemaSkill
- phase-2-conventionSkill
- phase-3-mockupSkill
- phase-4-apiSkill
- phase-5-design-systemSkill
- phase-6-ui-integrationSkill
- phase-7-seo-securitySkill
- phase-8-reviewSkill
- phase-9-deploymentSkill
- plan-plusSkill
- starterSkill
- zero-script-qaSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.