agentleFS
EastkapEastkap/autocomp/.claude/skills/launch-directories/SKILL.md

launch-directories

Skill4 starsChanged 5 months agoMITView on GitHub

Launch a product across startup/launch directories (Product Hunt-style listings) by driving the real logged-in boseclaw Camoufox session — one tested, reusable flow script per directory, with proof screenshots + a live-status audit. Use when the user says "launch <product> on the directories", "submit to directories", "run the directory signups", "do the directory launch", or wants to add a directory, re-run a launch, or audit which listings went live.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Test with --isolated until the proof screenshot shows a real confirmation, then audit.mjs.

    line 64

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

launch-directories — submit a product to launch directories, with proof + audit

Everything lives in directories/. The model is: the loop (Claude) drives the captured boseclaw Camoufox session through each directory's submit flow — no API-keyed daemon, no fresh logins (boseclaw is already signed into Google + GitHub; the session accumulates each new site's cookies so re-runs are "hot login"). Each directory is a script; each run leaves proof; an audit re-checks that listings actually went live. This is our repeatable GTM muscle — we run it for every product we ship (Principle 17, distribution-first).

Layout

  • engine.mjs — Camoufox driver + action helpers H (goto/wait/waitFor/fill/type/click/clickIf/ select/upload/press/url/text/has/shot) + proof capture + ledger. Pre-downloads kit.logo → kit.logoFile.
  • flows/<directory>.mjs — one proven flow per directory: export const meta = {directory, home, submitUrl} and export default async function launch({H, kit}) { … return {status, listingUrl, notes} }. Statuses: submitted | already-live | queued | blocked | failed. Read flows/peerpush.mjs (OAuth + AI-autofill + logo upload + free-queue + decline-upsell), flows/awesomeindie.mjs (OAuth + form + tag-select), flows/pitchwall.mjs (dedup + AI-generate) as the templates.
  • kits/<product>.json — the reusable product data (name, url, tagline, summary, description, categories, pricing, logo, positioningRule). One per product; every flow reads from it.
  • launch.mjs — runner. audit.mjs — live re-check. merge-rows.mjs — fold parallel --isolated runs.
  • proof/ — submissions.json ledger + <slug>-<directory>.png proof shots + audit-*.png.

Verbs

Launch a product on one directory (or all):

cd directories
node launch.mjs peerpush --kit kits/weekly-brief.json          # one directory
node launch.mjs --all    --kit kits/weekly-brief.json          # every flow, sequentially

Loads the boseclaw session, runs the flow, saves the session back (hot login persists), writes a proof screenshot + a ledger row. --headful to watch it; --state <file> for another identity.

Audit which listings are actually live (run daily / from a sub-agent):

node audit.mjs                       # re-check every ledger row, fresh screenshots
node audit.mjs --slug weekly-brief   # just one product

Marks each 🟢 live (listing loads + name present) / 🟡 pending (queue/review wording) / 🔴 missing (404 or name absent → needs resubmit). Rewrites the ledger's audit field + timestamps. Note: directories with no public per-product URL until approval (e.g. awesomeindie, hot100) read as missing while pending — eyeball those from the proof shot rather than trusting the flag.

Parallelize a big batch (one subagent per directory): each subagent authors its flows/<dir>.mjs and runs node launch.mjs <dir> --kit … --isolated (writes proof/row-<slug>-<dir>.json, does NOT touch the shared ledger/session → no races). Then merge: node merge-rows.mjs.

Adding a new directory (the "script per directory" step)

  1. Copy the closest template from flows/ (OAuth-form → awesomeindie; AI-autofill+logo → peerpush; dedup+generate → pitchwall).
  2. Rewrite the selectors/steps for the new site. Develop by dropping await H.shot('probe') steps and Reading the PNGs in proof/; dump form field selectors with node ../browser/camoufox.mjs act steps.json --state ../.secrets/boseclaw-state.json using a {"fields":true} step.
  3. Prefer OAuth ("Continue with Google/GitHub" → click text=Bose Claw → text=Continue): it usually skips the site's CAPTCHA entirely. Camoufox's real fingerprint passes Cloudflare Turnstile/interstitials passively; hCaptcha/reCAPTCHA image challenges may hard-block — then return blocked with proof, never fake a submission.
  4. Test with --isolated until the proof screenshot shows a real confirmation, then audit.mjs.

Rules (match the per-flow enforcement)

  • FREE tier only — never enter payment, always decline featured/expedite/paid upsells.
  • Positioning — obey kit.positioningRule; for Weekly Brief: it RANKS newsletters, it does NOT summarize — never write "summarize". Overwrite any AI-generated copy with kit.summary/kit.description.
  • Never invent traction — leave metrics blank unless measured.
  • Honest proof — a blocked/failed row with a screenshot is a real result; never record a submission that didn't happen (CLAUDE.md hard rules + principle 13).
  • Dedup — check for an existing listing first; return already-live instead of duplicating.

Bot identity, not gated

Directory submissions under the boseclaw bot identity are pre-authorized GTM (CLAUDE.md) — run them, don't ask. The gate is only for spend, the human's personal identity, and email-domain reputation.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.