Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add mackinleysmith/convex-worktree-isolation --skill setup-convex-worktreesgit clone --depth 1 https://github.com/mackinleysmith/convex-worktree-isolationWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mackinleysmith/convex-worktree-isolation/setup-convex-worktrees)<a href="https://agentmods.dev/skills/mackinleysmith/convex-worktree-isolation/setup-convex-worktrees"><img src="https://agentmods.dev/badge/skills/mackinleysmith/convex-worktree-isolation/setup-convex-worktrees.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00095 | $0.01509 |
| Opus 5 | $0.00048 | $0.00754 |
| Sonnet 5 | $0.00019 | $0.00302 |
| Haiku 4.5 | $0.00010 | $0.00151 |
Grade A, and why
setup-convex-worktrees scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 177 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Set Up Convex Worktrees
Give each linked Git worktree a short-lived Convex dev deployment, coherent baseline data, and an explicit agent invariant. Adapt the bundled implementation to the target repository instead of imposing its example layout or schema.
Workflow
1. Inspect before editing
- Read every applicable repository instruction file.
- Inspect the worktree state and preserve unrelated changes.
- Identify:
- the package manager and repository root;
- the Convex package version and directory;
- backend and frontend env files;
- the frontend's public Convex URL variable;
- existing dev scripts, seed functions, and worktree tooling;
- authentication, scheduled jobs, webhooks, and provider side effects.
- Read
convex/_generated/ai/guidelines.mdcompletely when present. If it is missing or stale, recommendnpx convex ai-files install. - Read references/adaptation.md before choosing paths, seed behavior, or a validation strategy.
Do not assume a monorepo. Do not replace compatible existing tooling merely to match the templates.
2. Choose the installation shape
Use the smallest shape that solves the collision:
- Always isolate the Convex dev deployment.
- Sync a separate client env file only when the Convex CLI does not already write the frontend's public URL variable.
- Add port isolation only when multiple frontend dev servers will run.
- Mirror only non-secret, checkout-specific values to the deployment.
- Keep remote tunnels, process supervisors, and hosting previews outside this skill unless the user explicitly includes them.
Require Convex 1.34.0 or later for deployment selection by reference. Prefer the repository's installed Convex CLI over a global version.
3. Install the provisioner
For a new installation, run the bundled no-overwrite copier:
node <skill-dir>/scripts/copy-template.mjs --target <repository-root>
Add --next only for a Next.js app that needs .env.local PORT support:
What ships with it
9 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 254 B
- assets/template/AGENTS.md.snippet 1.3 KB
- assets/template/convex/schema.snippet.ts 727 B runs code
- assets/template/convex/seed.ts 2.1 KB runs code
- assets/template/scripts/next-dev.mjs 1.6 KB runs code
- assets/template/scripts/setup-worktree-convex.mjs 9.6 KB runs code
- assets/template/worktree-convex.config.mjs 969 B runs code
- references/adaptation.md 5.8 KB
- scripts/copy-template.mjs 2.1 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 7d ago First seen · 177 lines · 95 tokens per session scan A 3b4719a91dfd
setup-convex-worktrees is a skill published in the GitHub repository mackinleysmith/convex-worktree-isolation (2 stars, last pushed 1mo ago), licensed MIT. It adds 95 tokens to every session and 1,509 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
ship
Use when code is ready to ship — creates PRs, merges, deploys, and verifies. Handles the full PR-to-production pipeline. Triggers on /ship.
agentxflow-coordinator
Authoritative multi-agent engineering coordinator for parallel workflows with isolated Git worktrees, write scope locking, automated machine verification, and serialized FIFO merge queue.
shiploop
Self-improving multi-agent harness: wraps N git sub-repos as one workspace, dispatches named tickets to cheap-floor headless agents that escalate once on failure, promoting lessons into CLAUDE.md. Use when working in or scaffolding a meta-repo workspace (sub-folders each own .git; root has scripts/ + queue/tickets.md…
deploy-preview
Deploy a preview branch to the staging environment for review. Use when a feature branch is ready for visual review, when someone asks to "deploy this for review", or before opening a PR.
monitor-remote-jobs
Default workflow for durable local, SSH, Slurm, Apple-signing, or GitHub publishing work through RunBeacon. Use whenever Codex needs to call a remote host, securely save or select SSH/GitHub credentials, run or recover a long task, monitor training, publish commits, wait without model polling, cancel a remote process…
release-deploy
Prepare a public open-source project for release and deploy it to a detected target. Use at a release or submission boundary to verify setup, docs, secrets, tests, lint, build, and licenses, then follow the target-specific deployment path. Runs expensive checks sequentially. Never publishes, tags, commits, pushes, or…