Pre-built agent setups you can fork into a fresh Firecracker microVM. Forks capture the original creator's /persist — config, conversation memory, custom skills, datasets. Forking creates a new paid microVM ($25+/mo, 30-day money-back).
53 public templates · Collections · GTM tools · Read the guide
Claude Code
Runs gstack's /review, /qa and /cso on a schedule and writes findings to disk — then stops at /ship and prints the commands for you to run. The push is blocked in settings.json, not just discouraged.
Claude Code
Plans content from what you actually sell and what buyers ask — then halves your stated capacity so you finish it. Every entry names its product. Never posts.
Claude Code
Reports burn rate over short and long windows with a projected exhaustion date — and flags the green SLOs too loose to constrain anything. Never moves a target.
Claude Code
Checks the thing generic tools miss — every product lives at two URLs, and a broken canonical splits its signals. Paste-in fixes. Never edits your store.
Claude Code
Diffs two OpenAPI versions and gets the request-vs-response asymmetry right — each finding says which client breaks and how. Never blocks a PR.
Claude Code
Reviews a PR against your repo's own recorded decisions — boundary violations, dependency direction, reversed ADRs. No rule to cite means a question, not a finding.
Claude Code
Computes contribution, breakeven CPA and the conversion rate you'd need before any product opinion — and refuses a verdict without a real CPC.
Claude Code
Benefit-led descriptions that mark missing specs instead of inventing them — and include the who-it's-not-for section that cuts returns. Never publishes.
Claude Code
Audits your Squarespace SEO Titles, metas, empty product descriptions and duplicates from the public sitemap — with paste-in fixes. Never edits the site.
Claude Code
Screens product candidates on review moat, rating ceiling and the specific fixable complaints — and refuses to invent a sales number. Never sources.
Claude Code
Surfaces the prior-year safe harbour most owners have never heard of, breaks out SE tax, and never prints a due date as fact. Not tax advice.
Claude Code
Pulls the promises out of your sent mail, flags the ones carried for weeks, and plans 60% of real capacity — leading with what to renegotiate. Never schedules.
Claude Code
Ranks accessibility findings by who is blocked, not by count — with selectors, fixes, and an honest list of what automation can't check. Never certifies.
Claude Code
Aggregates a year of contractor payments per payee, applies the card and corporation carve-outs, and leads with your missing W-9s. Not tax advice.
Claude Code
Reorders your enrichment waterfall by cost-per-incremental-hit and finds the providers billing for rows already filled. Never spends a credit.
Claude Code
Tests every Linear bug against the four things that make it reproducible, then drafts the exact questions for whatever is missing. Never closes an issue.
Claude Code
Drafts eBay listings that fit the 80-char title cap, fill every item specific, and price from sold comps — not asking prices. Never lists.
Claude Code
Finds the Sponsored Products spend that never converted, then drafts harvest, negate and bid-down changes with the arithmetic shown. Never changes bids.
Claude Code
True margin from landed cost — not markup, not invoice cost — plus price scenarios with the break-even volume change. Never changes prices.
Claude Code
Tracks competitor prices on landed cost with variant and stock context, and reports movement against recorded history. Never changes your prices.
Claude Code
Reconciles bank, processor, AR and AP exports into a close packet with tie-outs shown. Unexplained variance is a finding — it never plugs. Never posts entries.
Claude Code
Ranks Datadog monitors by action rate, not volume — then proposes specific thresholds with the percentile behind them. Never mutes a monitor.
Claude Code
Sorts a Dependabot backlog into routine, needs-migration and security — then drafts the groups and ignore rules that stop it regenerating. Never merges.
Claude Code
Triages GitHub secret-scanning alerts by blast radius into a rotate-first runbook. Never closes an alert, never prints the secret.
Claude Code
Ranks accounts by change from their own baseline into a HIGH/MEDIUM/WATCH triage list with the question to ask. No fake probabilities. Never contacts customers.
Claude Code
When main goes red, draft the smallest green fix PR. Human merge gate. Never force-push main.
Claude Code
Sequence, contracts, and data model deltas before sprawl. Lights on for long-lived decisions. Pair with program design.
Claude Code
Directional maintainability smells on a PR — complexity, duplication, shotgun risk. Draft review only. Tests green ≠ maintainable.
Claude Code
Problem, success metric, and rough HTML mockups before any code. Mockups are cheaper than slop PRs.
Claude Code
Large PRs get a logical diff tree — ordered by change intent, not alpha. Review energy heuristic. You still read the code.
Ready to fork
Fork a public template into a fresh microVM, drop in your model keys, and put it to work. 30-day money-back on first purchase · no free trial · from $25/mo.