agentsclimarketplace
Sign in

Init

Skill mk668a/FDE-SKILLS/skills/init

Use when the user wants to set up an FDE-SKILLS workspace or start a new client engagement (e.g. "set up fde", "scaffold fde here", "new engagement with Acme", "start the Globex engagement"). First run creates the .fde/ deterministic spine (config, schemas, bin scripts, empty corpus); with a client name it also creates the per-client engagement directory.From its SKILL.md

Install
npx -y skills add mk668a/FDE-SKILLS --skill init

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.8 KB, 632 tokens by cl100k_base, as published. Nobody here has run it

fde-init — scaffold the workspace and open engagements

Creates the .fde/ spine in the current directory (once), and one directory per client engagement (any time). A workspace is personal and spans all clients so knowledge compounds across them; isolation between clients is by directory.

Steps

  1. Scaffold the spine — skip this whole step if .fde/config.yml already exists (don't clobber; say it's already initialized). The templates ride inside this skill, so ${CLAUDE_SKILL_DIR} points at them no matter how the pack was installed: 
    src="${CLAUDE_SKILL_DIR}/templates"
mkdir -p .fde/bin .fde/schemas .fde/corpus engagements
cp "$src/config.yml"           .fde/config.yml
cp "$src"/schemas/*.schema.yml .fde/schemas/
cp "$src"/bin/*.sh             .fde/bin/
chmod +x .fde/bin/*.sh
  2. Gitignore the confidential layer — if the workspace is a git repo, add to .gitignore: 
    engagements/
.fde/dashboard.html
    Everything under engagements/ is raw client material and must never be committed; the dashboard contains client slugs. The shared .fde/corpus/ (anonymized) and .fde/schemas/ are commit-safe.
  3. Open the engagement — if the user named a client: 
    .fde/bin/fde-new.sh "<Client Display Name>"
    This creates engagements/<slug>/{notes,deliverables}/, writes engagement.yml, and registers the slug in .fde/index.yml.
  4. Print the layout and point at /fde:draft for the first deliverable.

The workspace layout

.fde/
  config.yml      deliverable types + paving threshold + extra identifiers
  schemas/        <type>.schema.yml — slot defs that evolve across engagements
  corpus/         <type>.yml — anonymized priors with sighting counters
  bin/            deterministic scripts (no LLM): new, coverage, promote,
                  validate, identifiers, dashboard
  index.yml       registry of engagements (created by fde-new.sh)
  dashboard.html  the compounding wall (fde-dashboard.sh; never commit)
engagements/
  <client>/       one dir per client — CONFIDENTIAL, never cross-read
    engagement.yml, notes/, deliverables/

Hard rules

  • Never write client-identifying content into .fde/corpus/. Only /fde:promote moves knowledge up, and it anonymizes on the way.
  • The spine scripts are the source of truth for counting/redaction/paving. Don't reimplement their logic in prose — call them.

What ships with it: 15 files

44.7 KB alongside SKILL.md, 6 of them executable

templates/

Keep looking

Skills are one crate of 326,629. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.