builder-skills
AGENTS.md404 starsChanged 45 days agoApache-2.0View on GitHub
Agent skills for builders shipping Convex apps. This file is context for agents working inside this repo. It also symlinks to CLAUDE.md. Seventeen skills in skills/<name>/. Fourteen cover Convex. Three cover how to run a project: PRD first, docs synced from git evidence, git commands that cannot destroy work. Each skill is a folder with SKILL.md, an agents/openai.yaml for Codex icons, and sometimes a references/ folder that the skill loads on demand. The CLI in bin/cli.js installs skills into .claude/skills,…
1 other project uses a version of this.
Adopted by
Most recently taken up by
- waynesutton/convexskillsword for word, first seen Sep 28, 2026
Other versions1 project uses a version of this.
Word for word · 1
- waynesutton/convexskills★ 405100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
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 31,774 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 repository19 other files it gives its agents.
- templates/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- avoid-feature-creepSkill
- convex-agentsSkill
- convex-best-practicesSkill
- convex-component-authoringSkill
- convex-cron-jobsSkill
- convex-file-storageSkill
- convex-functionsSkill
- convex-http-actionsSkill
- convex-migrationsSkill
- convex-realtimeSkill
- convex-schema-validatorSkill
- convex-security-auditSkill
- convex-security-checkSkill
- convexSkill
- git-safetySkill
- project-docsSkill
- project-workflowSkill
The file
We split this AGENTS.md into the rules it gives an agent
5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 21
No emojis. No em dashes. No marketing words.
- Very specific: line 23
Run npm run check before calling anything done.
- Very specific: line 28
After a change, sync task.md, changelog.md, files.md.
- Very specific: line 29
Before any git command that could discard work, see skills/git-safety.
- Distinctive: line 30
Do not commit, push, publish, tag, or deploy.
≈ “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
builder-skills
Agent skills for builders shipping Convex apps. This file is context for agents working inside this repo. It also symlinks to CLAUDE.md.
What this repo is
Seventeen skills in skills/<name>/. Fourteen cover Convex. Three cover how to run a project: PRD first, docs synced from git evidence, git commands that cannot destroy work. Each skill is a folder with SKILL.md, an agents/openai.yaml for Codex icons, and sometimes a references/ folder that the skill loads on demand.
The CLI in bin/cli.js installs skills into .claude/skills, .codex/skills, or .agents/skills. Templates in templates/ are starters a user edits. scripts/check-skills.mjs is the lint.
Docs first
For Convex APIs, fetch https://docs.convex.dev/llms.txt and follow the link for the area. Do not rely on memory for method names, validator names, or component APIs. When a skill cites an API, it was checked against that index; keep it that way.
Working on skills
- Frontmatter is
nameanddescriptiononly.namematches the folder. descriptionis third person, says what the skill does, and includes aUse when ...sentence.SKILL.mdunder 300 lines. Hard limit 500. Deep material goes inreferences/*.mdand gets linked from the body.- Code samples use the object form with
argsandreturnsvalidators,withIndexoverfilter,internal.*for scheduling. - No emojis. No em dashes. No marketing words.
- Adding a skill means updating
bin/cli.js,index.js,.claude-plugin/plugin.json, the README table,docs.md, and adding a symlink in.codex/skills/. - Run
npm run checkbefore calling anything done. It has to pass.
Working on this repo
- Non trivial work gets a PRD in
prds/. Seeskills/project-workflow. - After a change, sync
task.md,changelog.md,files.md. Seeskills/project-docs. Dates come fromgit log --date=short. - Before any git command that could discard work, see
skills/git-safety. - Do not commit, push, publish, tag, or deploy. Print a suggested commit message and stop.
Layout
skills/<name>/SKILL.md the skill
skills/<name>/references/*.md loaded on demand from SKILL.md
skills/<name>/agents/openai.yaml Codex icon metadata
bin/cli.js builder-skills CLI
index.js programmatic API
scripts/check-skills.mjs lint, run with npm run check
templates/ starters installed by install-templates
.claude-plugin/ plugin.json and marketplace.json for Claude Code
.codex/skills/ symlinks into skills/ for Codex discovery
command/convex.md OpenCode slash command
prds/ PRDs for this repo, lessons.md, archive/Skills
| Skill | Load when |
convex | Convex task with no closer match |
convex-best-practices | reviewing patterns, OCC conflicts, ESLint setup |
convex-functions | writing any function in convex/ |
convex-schema-validator | tables, validators, indexes |
convex-realtime | frontend subscriptions, optimistic updates |
convex-http-actions | webhooks, REST, CORS |
convex-file-storage | uploads, serving, metadata |
convex-cron-jobs | crons, scheduled functions |
convex-migrations | live schema changes, backfills |
convex-agents | AI agents, tools, RAG |
convex-component-authoring | building a component |
convex-security-check | quick pass before merge |
convex-security-audit | full review before launch |
avoid-feature-creep | scope is drifting |
project-workflow | multi step work, PRDs, task.md |
project-docs | syncing changelog, files.md, task.md |
git-safety | any git command that could discard work |
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.