agentleFS
Lucas-GrilliLucas-Grilli/context-os-plugin/skills/project-kickoff/SKILL.md

project-kickoff

Skill1 starsChanged 3 months agoMITView on GitHub

Opens a new project in the second brain — folder, CLAUDE.md, memory.md, row in the "Active Projects" table and entry in the index, so it is reachable from the Router straight away. Use for "open a new project", "project kick-off", or whenever a new line of work is born in 10_projects/.

14 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 13 edited versions we read keep these.

  • 2

    Distinctive

    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:

    archive.md — only when needed: Do not create it at kick-off: it is born the first time a state in memory.md is replaced, not added to.

    line 22
  2. Distinctive:

    If the project is born by moving or merging content that already exists elsewhere, scaffolding is not enough: also apply "always links, …

    line 27
  3. Distinctive:

    If it is not clear in one sentence what the new project does, ask before proceeding.

    line 28

“Common” is measured: kept by at least 6 in 10 of the 13 edited versions of this file we could read. “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

Project kickoff

v2.0

Goal: stop projects being born without a CLAUDE.md/memory.md, or without a row in the root table — the two most frequent holes.

Steps

  1. Check it does not already exist — grep/glob on the name, synonyms, domain. If something close already covers the same need, propose that before creating from scratch.
  2. Weigh it — one question: will this project accumulate decisions and state over time that you will need to re-read a week from now?
    • No (a routine producing repeated output, no decisions to re-read) → skip steps 3-4, an operational file with 2 lines at the top is enough. If it is that light, consider whether it is a procedure rather than a project at all.
    • Yes (clients, live channels, decisions that change) → full scheme.
  3. Folder + CLAUDE.md — copy 99_system/templates/claude-template.md to 10_projects/<name>/CLAUDE.md: role in one line, scope (what belongs, what does not and where the rest goes), internal structure only if it is not obvious, rules inherited from root made explicit.
  4. memory.md — copy 99_system/templates/memory-template.md to 10_projects/<name>/memory.md, fill in the header, delete the sections that stay empty.
  5. Row in the "Active Projects" table of the root memory.md — correct path, status "Active", next step in one sentence. Same session, it is a macro event. This applies to the light projects of step 2 too.
  6. Entry in 10_projects/projects_index.md — same operation. Without it the project exists on disk but is unreachable through the cascade (Rule 1). Two distinct writes: the table carries the state, the index carries reachability.
  7. archive.md — only when needed. Do not create it at kick-off: it is born the first time a state in memory.md is replaced, not added to. It does not apply to light projects.

Hard rules

  • Kick-off always in the root chat, never inside a project chat.
  • If the project is born by moving or merging content that already exists elsewhere, scaffolding is not enough: also apply "always links, never free-text paths" and check (grep) every reference to the old path.
  • If it is not clear in one sentence what the new project does, ask before proceeding.

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.