One install gives your agent a discovery setup for a GitHub repo, skill, plugin or MCP server: it audits where you rank today and who appears instead, derives the queries strangers actually type, then drafts the About text, topics, README opening, release notes, site page and directory plan — and books a dated re-check to see what moved.
install
Detects Claude Code, Cursor, Codex, Qoder and more, then writes the skill into the right folder. Add --dry-run to see the plan first.
npx skills add arm-pz/findable --skill findable
the kit
One audit produces one file: a discovery-kit.md that records what is true today, says what should change, and hands you the commands. It writes for people and search engines; the SKILL.md that triggers the agent is a separate job.
Where the project appears today, and where comparable projects do.
What a stranger types. At least two per capability, none needing your name.
Only when something else already holds the name.
The surfaces GitHub and directories actually read.
Directories worth checking, with each one's submission rules read first.
The same queries, re-run against the recorded baseline two to three weeks later. Indexing lags, so a check the day after publishing proves nothing.
ground rules
It never claims stars, installs, rankings, benchmarks or compatibility the repo does not support. Every figure on this page was fetched from GitHub in the run it describes, and the date of that fetch is stated beside it.
The kit ends in commands you run: gh repo edit, gh release create, submission steps. The tool drafts them and stops. No description, topic, release or page changes behind your back.
No star trading or buying, no fake reviews, no topic stuffing, no undisclosed promotion, no asking for stars in communities. It also leaves example slots empty and labelled rather than writing sample output a project never produced.
Free, open, and it only ever drafts. You run the commands.
Install findable Read the source