craft
AGENTS.md0 starsChanged 56 days agoMITView on GitHub
This repository is a skills library and reference-only router for upstream packs. See craft.manifest.yaml. Install: npx skills add addyosmani/agent-skills. When adding or refining skills, read using-craft first.
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 31,542 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 repository5 other files it gives its agents.
- craft-adoptSkill
- craft-promoteSkill
- engineering-ledgerSkill
- product-managementSkill
- using-craftSkill
The file
We split this AGENTS.md into the rules it gives an agent
5 rules. About 0 are common, 1 is distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 7
Read README.md and docs/lifecycle-map.md.
- Very specific: line 25
When adding or refining skills, read using-craft first.
- Very specific: line 31
Mandatory Skill-First Routing: Every discussion, architecture design, development, and debugging task must flow through the corresponding …
- Very specific: line 39
Anti-Drift: Use the single best skill for the active phase (via using-craft).
- Very specific: line 40
Phase Transition Thresholds: Never jump phases without satisfying the previous phase's exit gate evidence in phases.md.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 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
Craft — Agent Instructions
This repository is a skills library and reference-only router for upstream packs.
Working in this repo
- Read README.md and docs/lifecycle-map.md.
- Each skill:
skills/<name>/SKILL.mdwith YAML frontmatter (name,description). - Never copy upstream Addy/Superpowers SKILL.md bodies into craft — update router tables instead.
- New Craft-native skills: follow skills/using-craft/SKILL.md quality bar.
Craft-native skills (v0.1)
| Skill | Role |
using-craft | Phase router → upstream skill names |
engineering-ledger | Project ledger read/write loops |
craft-adopt | Bootstrap target projects |
craft-promote | Universal LL → new skill |
Upstream (referenced, not vendored)
See craft.manifest.yaml. Install: npx skills add addyosmani/agent-skills.
When adding or refining skills, read using-craft first.
Craft (Orchestration, Skills & Ledger)
- Read Ledger First: Non-trivial work must read
docs/engineering-ledger/INDEX.mdand active phase inphases.mdbefore taking action. - Mandatory Skill-First Routing: Every discussion, architecture design, development, and debugging task must flow through the corresponding skill in
.agents/skills/. Before executing any phase, cross-check installed skills and load the best domain skill:- Ideation & Discovery: Load
idea-refineorinterview-meto sharpen ambiguous requirements. - System Architecture & Design: Load
system-design(stress-test failure modes, trade-offs, scaling, data models) along with building blocks (api-design,architecture-diagram,caching,resilience-failure,data-storage,messaging-streaming). - Specification & Decisions: Load
spec-driven-development. For irreversible forks, create an ADR indocs/decisions/perdocumentation-and-adrs. - Implementation & TDD: Load
incremental-implementationfor thin slices andtest-driven-development(Red-Green-Refactor). - Debugging & Errors: Load
debugging-and-error-recovery(systematic root cause analysis, never guess). - Verification & Quality: Load
constraint-driven-developmentandcode-review-and-qualitybefore completing features. - UI / Frontend: Load
ui-ux-pro-maxandfrontend-ui-engineering.
- Ideation & Discovery: Load
- Anti-Drift: Use the single best skill for the active phase (via
using-craft). Never bypass skills or invent ad-hoc processes when an established skill exists. - Phase Transition Thresholds: Never jump phases without satisfying the previous phase's exit gate evidence in
phases.md. Anti-drift thresholds must be verified before proceeding to prevent premature implementation. - Market Validation Gate: In
docs/northstar.md, evaluate the Market Validation Score (0–10). A score of >= 7/10 is required to enterBuild. If < 7, stop or pivot. - Durable Ledger Updates: Append tactical decisions (
DR-###) and lessons (LL-###) todocs/engineering-ledger/, and updateINDEX.mdbefore ending substantive sessions.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
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.

