skills/help/SKILL.md
--- tldr: List all eidos skills with descriptions category: core --- # /eidos:help List available eidos commands, explain the workflow, and offer interactive guidance. ## Usage ``` /eidos:help ``` ## Instructions ### 1. List Skills Read all SKILL.md files in `${CLAUDE_PLUGIN_ROOT}/skills/`. For each, extract `tldr` and `category` from YAML frontmatter. Group by `category` in this order: core, planning, observation, utility. Present as a formatted list: ``` Eidos — spec-driven developmen
npx skillsauth add agenticnotetaking/eidos skills/helpInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
List available eidos commands, explain the workflow, and offer interactive guidance.
/eidos:help
Read all SKILL.md files in ${CLAUDE_PLUGIN_ROOT}/skills/.
For each, extract tldr and category from YAML frontmatter.
Group by category in this order: core, planning, observation, utility.
Present as a formatted list:
Eidos — spec-driven development for Claude Code
Core Loop:
/eidos:<name> <tldr>
...
Planning:
...
Observation:
...
Utility:
...
After the skill list, always show a brief workflow overview:
How it works:
eidos/ what it SHOULD be (specs, intent)
memory/ how we got here (plans, decisions, research)
src/ what it IS (code)
Eidos keeps specs and code in sync bidirectionally:
Spec → Code Write what you want, then make it real (/push)
Code → Spec Extract intent from existing code (/pull)
The Loop Edit spec → push → test → adjust spec → repeat
Specs are markdown files with wiki links forming a relationship graph.
Leave {{comments}} in specs for AI-assisted refinement (/refine).
Plans in memory/ persist across sessions and track multi-step work.
After the overview, ask:
Want help figuring out where to start?
If the user declines, stop here.
If the user accepts, have a short interactive conversation to understand their situation:
eidos/? Any plans in memory/?Based on their answers, suggest a concrete starting path with specific skills.
New project, starting from ideas:
Dump your notes into
eidos/seed.md— doesn't have to be clean. Then/eidos:specturns them into a proper spec. Review it, leave{{comments}}on anything that's off, run/eidos:refine. When you're ready to build:/eidos:planfor big work,/eidos:pushfor direct implementation.
Existing project, no specs yet:
/eidos:pullextracts a spec from your code — pick a feature or component to start with. Read the generated spec, leave{{comments}},/eidos:refineto iterate. Now you have a spec you can evolve. Edit it,/eidos:pushthe changes back.
Mid-project, specs exist:
/eidos:driftshows where specs and code have diverged./eidos:nextshows open items across plans, todos, and specs./eidos:coherencechecks specs against each other for contradictions.
Just exploring:
Read [[spec - eidos - spec driven development loops]] for the full design. Try
/eidos:pullon a small piece of code to see what a spec looks like. The workflow clicks once you've done one push-pull cycle.
These are examples — adapt to what the user actually tells you. The goal is a short, useful conversation that ends with a clear next action.
development
--- tldr: Check the status of messages you sent to peer inboxes category: core --- # /eidos:outbox Find messages **you sent** to peer repos and report their status. Your sent messages live in the peers' inboxes, not here. So this skill sweeps every registered peer and reads back what became of them. This is the mirror of [[spec - inbox skill - read and act on your own inbox]]. - The inbox is messages others sent **to** you. You own and edit it. - The outbox is messages you sent **to** others.
development
--- tldr: Send a message to a peer repo's inbox category: core --- # /eidos:message Write a message into another repo's `agent-inbox/`. The recipient is a peer repo, addressed by its registered name. See [[spec - agent communication - cross repo inboxes with append only messages and registry resolution]]. ## Usage ``` /eidos:message [recipient repo] [what to say] ``` ## Instructions ### 1. Resolve the recipient - Identify the recipient repo name. Ask if not given. - Resolve its path: `
development
--- tldr: Read and act on your own inbox category: core --- # /eidos:inbox List and work through messages other repos have left in your `agent-inbox/`. You own this inbox. You may edit its files freely. Agents from other repos may push **new** messages into your inbox at any time. They never edit existing files. Only you do. So re-check the inbox. Don't assume it is unchanged since you last looked. See [[c - foreign agents append to an inbox - only the owner edits in place]]. See [[spec - ag
tools
--- tldr: Move done files into archive subdirs (memory/ and agent-inbox/) category: utility --- # /eidos:archive Retire finished files into the `archive/` sibling of their directory. Done `memory/` files into `memory/archive/`, done inbox messages into `agent-inbox/archive/`. A `git mv`, never a rename. The basename is preserved, so wiki links keep resolving. See [[c - archiving moves done files into an archive subdir preserving basename and links]]. ## Usage ``` /eidos:archive [optional fil