Skip to content
Productivity
Skill

/idea

Use when capturing ONE new idea the user voices and wants recorded — "save this idea", "I have an idea", "log this idea", "/idea", "idea: ...". Creates a provenance-tracked folder (one folder per idea) in your ideas repo, dedups against an index, optionally mirrors to a GitHub

From plugin
personal-corp-os
24435 skills7 agents

Claude Code

$ npx -y skills add serejaris/personal-corp-os --skill idea --agent claude-code

SKILL.md

idea.SKILL.md
name: idea
description: >-
  Use when capturing ONE new idea the user voices and wants recorded — "save this idea",
  "I have an idea", "log this idea", "/idea", "idea: ...". Creates a provenance-tracked folder
  (one folder per idea) in your ideas repo, dedups against an index, optionally mirrors to a
  GitHub Project view filtered by label:idea. NOT for mining ideas out of transcripts/meetings
  (use a dedicated extraction skill) and NOT for dated work tasks (use an issue tracker / the
  manager skill). Triggers also in Russian: «запиши идею», «у меня идея», «зафиксируй идею».

Idea — fast capture of a single idea

Part of the Personal Corp framework — running a one-person business through AI agents.

One pipeline: your voice → a provenance-tracked folder in your ideas repo → (optional) an issue mirror on a GitHub Project view. The folder is the source of truth; everything else is a pointer to it.

This skill is **self-contained** — it inlines the folder structure, status enum, and provenance formats below. You don't need any external canon file.

What this skill is NOT for

  • **Mining ideas out of a transcript / meeting / thread** — that's bulk extraction from a source. Use a separate extraction skill for that. This skill captures ONE idea the user is voicing right now.
  • **A dated work task with a deadline** — use an issue tracker / the `manager` skill. Ideas have no deadline; they're seeds.

Setup

Before first use, define this in your project's `CLAUDE.md`:

## Idea Config

### Ideas repo path
Where idea folders live (one folder per idea):
- ideas_repo: ~/Projects/ideas

### Default author
Name written into `first_voiced_by` when the user doesn't say otherwise:
- default_author: Me

### GitHub mirror (optional)
Mirror each idea as an issue (added to a Project view filtered by `label:idea`):
- mirror_enabled: false
- owner: your-github-handle               # only if mirror_enabled: true
- ideas_repo_name: ideas                  # GitHub repo name for the issue, only if mirror_enabled: true
- project_number: <your Project number>   # only if mirror_enabled: true

(if mirror_enabled is false, the skill only writes to the ideas repo — no issue, no Project)

No separate init skill needed — this section is the setup. If `mirror_enabled` is false, the GitHub Project step is skipped entirely; the idea still lives as a folder.

One question — max

Ask **at most one** short question, and only if the idea's essence can't be extracted from the wording (what is it / for whom). Do NOT ask about mechanics, pricing, risks, or tech stack — capture the raw material as-is and write the gaps into `next-steps.md` as open questions. The question **"how should I record it — issue / file / both?"** is forbidden: the form is fixed by this skill. Never block on form.

Status enum (in `idea.md` frontmatter)

  • `seedling` — just voiced, not yet worked on
  • `exploring` — first research gathered, `market.md` filled
  • `building` — landing / pilot / prototype in progress
  • `live` — a public artifact exists
  • `parked` — set aside, reason in `next-steps.md`
  • `dropped` — closed, reason in `next-steps.md`

A freshly captured idea is always `seedling`.

Provenance formats (rows in `sources.md`)

Every idea needs at least one source. Each sighting is one row. Genericize the source kind — no real internal paths:

| Source kind | Format in `sources.md` | |---|---| | Meeting note | `meeting note — {who/where}, {date}` | | CRM card | `CRM card — {contact}, {date}` | | DM | `DM — {channel}, {date}` | | Voice draft (solo) | `voice draft, {date}` | | Reading / web | `{url}, {date}` | | Direct capture | `direct /idea invocation, session, {date}` |

Steps

1. **`date`** — resolve the real date. Between 00:00–04:00, "today" means the previous calendar day.

2. **Dedup against the index.** Read `INDEX.md` at the root of your ideas repo. Match is **semantic, not string** ("a board of advisors run on agents" ≈ `advisory-agent-grounded-data`). If unsure, open the candidate's `idea.md` to compare.

  • **Match found** → this is a repeat sighting, not a new idea:
  • append a row to the existing folder's `sources.md`,
  • in `idea.md` bump `last_touch` and increment `sources_count` (do NOT touch `first_voiced`),
  • if mirror is enabled, add a one-line comment to the issue mirror ("voiced again: {date}, {context}"),
  • regenerate `INDEX.md`, commit, report. **STOP.** Do not create a new folder.

3. **New idea** → create the folder `ideas/YYYY-MM-DD-{slug}/` with the template files below.

  • **Slug**: kebab-case, 3–5 words, by meaning not by form (`live-sales-coach-b2b`, not `b2b-product-1`). The date prefix is the date of **first capture** — it never changes on later sightings.
  • In `idea.md`: `title`, `status: seedling`, `first_voiced_by: {default_author from config unless user said otherwise}`, an "Essence" of 1–2 sentences, and the **raw verbatim wording** preserved exactly as voiced. Interpretation, if any, goes in a separate section — never overwrite the raw wording.
  • In `sources.md`: one row for this sighting (see provenance formats).
  • `market.md` is optional (leave the stub; fill only when researched).
  • `next-steps.md` holds open questions and the "what next?" — this is where pricing/mechanics/stack gaps land.

4. **Regenerate `INDEX.md`** (format below). It's auto-generated — don't hand-edit individual rows; rewrite the whole file from the folders.

5. **GitHub Project mirror — OPTIONAL** (only if `mirror_enabled: true` in config). The folder is the source of truth; the issue is just a pointer for board view. See "Project mirror" below.

6. **Commit + push** to the ideas repo: `feat: idea {slug}` (invoking this skill is consent to record; the ideas repo is private).

7. **Report** — one line per item: folder path, issue `#N` (if mirrored), and the Project view link (if mirrored).

Template files

Create these four files in `ideas/YYYY-MM-DD-{slug}/`.

Read more
@serejaris
Ships withpersonal-corp-os

Manager uses your local Manager Config for repositories and boards; the distributed skill contains no personal workspace logs.

24435 skills7 agents
Get the whole plugin

Stars on GitHub

Momentum

stars in total
244
Star history of serejaris/personal-corp-os on GitHub, a chart by star-history.com

Other skills on personal-corp-os.

daily-retro

daily-retro

Experimental. Daily retro from your agents' own logs: reads today's Claude Code and Codex sessions on this machine, groups them…

art-director

art-director

Orchestrate iterative visual style searches with branch prompts, decision graphs, feedback loops, and final direction selection.

avatar

avatar

Робот агента Personal Corp. Агент делает себе голову робота из набора деталей, хранит её в штабе в stack/avatars.json и кладёт в…

benchmark

benchmark

Use when choosing between engines or models for an agent pipeline and the answer must come from measurement on your own data, not…

More skills from other plugins.

i-have-adhd

i-have-adhd

Shape output for a reader with ADHD: lead with the next action, number multi-step work, restate state across turns, suppress…

@ayghri@ayghriInstall Skill
ab-test-analysis

ab-test-analysis

Analyze A/B test results with statistical significance, sample size validation, confidence intervals, and ship/extend/stop…

@phuryn@phurynInstall Skill
distilly

distilly

Set up the owner profile so a personal agent knows who it works for (first-run setup, \"get to know me\"), or distill colleague,…

Questions about idea.

Which coding agents can use idea?
idea installs in Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and Windsurf with the skills CLI, using each agent's flag shown on this page. Flowy's auto-invocation, which fires a skill when your prompt matches, runs in Claude Code only.