Skip to content

mikehasa/golive-skill / skills/golive

golive

Flagged

Take an agent-written app from repo to live production on the user's OWN accounts, with providers they choose (hosting, database, auth, payments, email, domain/DNS). The human connects accounts and approves changes; supported wiring operations run through a local CLI and produce verification evidence with explicit limits. Use when the user wants to ship, deploy, go live, launch, publish, or put their app online, or asks to wire up env vars, webhooks, auth settings (signup, email confirmation, password policy), a real signup → confirmation email → login journey, password recovery, account isolation between two users, auth redirects, email DNS or a custom domain.

FileKindSize
LICENSEother1.0 KB
SKILL.mdskill md50.0 KB
THIRD_PARTY_NOTICES.mdmarkdown817 B
references/cloudflare-dns.mdmarkdown7.9 KB
references/godaddy.mdmarkdown3.9 KB
references/guided.mdmarkdown9.2 KB
references/neon.mdmarkdown4.4 KB
references/netlify.mdmarkdown6.9 KB
references/plan-and-verify.mdmarkdown47.4 KB
references/porkbun.mdmarkdown3.4 KB
references/posthog.mdmarkdown10.2 KB
references/resend.mdmarkdown8.6 KB
references/sentry.mdmarkdown10.7 KB
references/stripe.mdmarkdown13.1 KB
references/supabase.mdmarkdown49.7 KB
references/troubleshooting.mdmarkdown5.2 KB
references/updates.mdmarkdown5.0 KB
references/uptimerobot.mdmarkdown12.3 KB
references/vercel.mdmarkdown13.2 KB
release.jsonconfig2.4 KB
scripts/golive.mjsscript1.2 MB
scripts/install-cli.mjsscript5.0 KB
scripts/install-lib.mjsscript24.4 KB

references/troubleshooting.md

View on GitHub
# Fix a setup failure and return to the deployment

Use this reference when a command fails during the skill's normal flow. Preserve the user's app,
provider choices, exact project scope, current plan and completed step records. A troubleshooting
detour ends with a passing check or a scoped external observation establishing the repair, plus a
clear continuation point. For guided providers, record external evidence separately; CLI checks
that lack provider coverage remain skipped.

## Installed, but command not found

First distinguish a missing installation from a shell lookup problem. On macOS/Linux, narrow
diagnostics include `command -v node`, `command -v npm`, `command -v <vendor-cli>` and
`npm prefix -g`. For an npm-installed CLI, check whether its declared executable exists under that
prefix's `bin` directory and run that exact executable with `--version`. Use the actual package's
binary name; package names and commands do not always match. On Windows, use the shell's command
lookup and npm's platform-specific executable location instead of assuming a `bin` subdirectory.

The human's separate Terminal and the agent may use different Node installations or PATH values.
Verify both contexts when necessary. An export in the human's terminal does not alter the already
running agent. A new terminal or `hash -r` can refresh shell lookup, but neither adds a missing
directory to PATH.

If the binary exists and works, correct the confirmed PATH/executable lookup with the smallest
appropriate change. Explain any persistent shell configuration edit; preserve existing entries
and executable links. Use a temporary PATH adjustment for a diagnostic when sufficient. Do not
hard-code a previous user's home directory, overwrite another CLI, switch Node installations,
repeatedly reinstall, or introduce `sudo` merely because command lookup failed. If the binary is
absent, follow the selected provider's supported installation method, then recheck discovery.

Only request non-secret diagnostic output from the human if the agent cannot inspect the relevant
terminal context. Never request a whole environment dump, npm config, shell startup file, auth
file or command history: those can contain credentials.

## Login or provider operation failed

Once the executable works, run the skill's `doctor` for the selected stack. If login is genuinely
missing/expired, have the human complete the provider's browser login in their separate terminal,
then rerun the account check. A login failure does not imply they need a second manual API token;
follow the adapter's supported path and describe any actual limitation.

A successful login followed by a failed create/deploy may indicate input shape, permissions,
Free quota, rate limits or a provider/CLI change. Do not prescribe reinstalling or generating a
new token without evidence. Use bundled provider references first; consult current official docs
for a concrete unresolved mismatch. Keep raw provider errors out of chat when they may carry
credentials; inspect only safe status codes, field names and projected resource metadata.

After an ambiguous write failure, inspect the exact approved scope and resource identity read-only
before retrying. A missing local step record does not prove the provider made no resource. Never
create a second project as a retry, adopt a same-named project without approval, weaken credential
handling, or bypass a plan guard to get past the error.

## Native credential entry needs attention

Read only the command's status metadata; never the saved file or raw dialog output. `cancelled`
means the human stopped: wait rather than reopening it. `unsupported-platform` or
`dialog-unavailable` can use the private-file editor fallback. A timeout needs the human to be
ready before another attempt. An invalid value was not saved; explain the supported single-line
input without quoting what they entered. `already-exists` needs an intentional replacement choice.

For `unsafe-path`, `concurrent-change` or `write-failed`, inspect ownership, permissions, link and
lock metadata, resolve that specific problem and preserve existing contents. Do not bypass the
checks with another writer or repeatedly request a new key. A `saved` result with `cleanupRequired`
means the key was saved but local staging/lock cleanup needs attention; do not ask for it again.
If `envOverride` is true, the existing process environment still wins: resolve which credential
source the human intends without reading back either value. Then resume the provider check.

## Resume the current stage

Read the current config, non-secret state and latest plan/result. Preserve completed steps and
recheck only the prerequisite affected by the repair. Re-plan if required; present any changed
scope or writes and obtain the required approval before apply. A prior approval remains useful
context but is not permission for new resources or changed destinations. Respect any stricter
per-apply approval rule set by the user.

Tell the human what the repair verified, what is already complete, and what deployment step comes
next. Continue within the existing authorization instead of making them repeat signup, provider
selection or successful logins. Keep unresolved checks open; a suggested fix alone is not success.
writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deploymentStars: 322kUpdated Claude CodeCodexVerifiedInstall
xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release.Stars: 118kUpdated Claude CodeCodexVerifiedInstall