/idea — Submit a New Product Idea

Capture a new product, feature, or internal-tool idea so it lands somewhere durable instead of evaporating in chat. This skill is intentionally lightweight: it adds an entry to the ideas backlog and (optionally) creates a tracking GitHub Issue. It does not replace /write-spec — that comes later, after the idea has been triaged.

Path resolution

Read the registry path via portfolio_registry, the per-project docs dir via portfolio_projects_dir, and the ideas backlog via portfolio_ideas_backlog — all from .claude/hooks/_lib-portfolio-paths.sh. Source the helper at the top of any bash block that touches those paths:

source "$(git rev-parse --show-toplevel)/.claude/hooks/_lib-read-config.sh"
source "$(git rev-parse --show-toplevel)/.claude/hooks/_lib-portfolio-paths.sh"
registry=$(portfolio_registry)

Defaults match today's single-fork layout (./apexyard.projects.yaml, ./projects, ./projects/ideas-backlog.md). Adopters in split-portfolio mode override the portfolio.{registry, projects_dir, ideas_backlog} keys in .claude/project-config.json. Don't hardcode literal apexyard.projects.yaml or projects/ paths in bash blocks — the helper resolves whichever mode the adopter is in. See docs/multi-project.md.

Usage

/idea Auto-tag inbound emails by intent
/idea Internal CLI for resetting staging data
/idea New product: AI design system linter

Where the entry goes

Every idea lands in projects/ideas-backlog.md at the root of your ops repo (your fork of apexyard). One shared backlog for every project — triage decides which project ends up owning a given idea.

If the file doesn't exist yet, create it with a header and a table.

Process

0. Write the active-issue-skill marker (REQUIRED — me2resh/apexyard#268)

Before any gh issue create (or other tracker CLI), write this skill's name to the active-issue-skill marker so require-skill-for-issue-create.sh lets the command through. At skill entry:

# Resolve the ops-fork root the SAME way the hooks do (_lib-ops-root.sh):
# anchor on the .apexyard-fork marker (split-portfolio v2 — onboarding.yaml
# lives in the sibling portfolio repo, NOT the ops fork), falling back to the
# onboarding.yaml + apexyard.projects.yaml pair (single-fork v1).
ops_root="$PWD"; r="$PWD"
while [ "$r" != / ]; do
  if [ -f "$r/.apexyard-fork" ] || { [ -f "$r/onboarding.yaml" ] && [ -f "$r/apexyard.projects.yaml" ]; }; then
    ops_root="$r"; break
  fi
  r=${r%/*}
done
mkdir -p "$ops_root/.claude/session"
echo "idea" > "$ops_root/.claude/session/active-issue-skill"

Remove the marker on every exit path (success, early-exit, user cancel, error):

rm -f "$ops_root/.claude/session/active-issue-skill"

The clear-issue-skill-marker.sh SessionStart hook sweeps stale markers from killed sessions, but a clean exit should never leave one behind. See AgDR-0030.

1. Parse the title

Take the title from $ARGUMENTS. If empty, ask:

What's the idea? Give me a short title (1 line).

2. Gather metadata

Ask conversationally (one question at a time, don't batch):

Category — must be one of four values. Present numbered options and re-prompt on invalid input:

Category?
  1. New Product
  2. Feature
  3. Internal Tool
  4. Process
>

Accept 1, 2, 3, 4 or the corresponding word (case-insensitive). On anything else, say Please pick 1-4 or type the category name. and re-ask. Loop until valid — do not silently accept garbage.

Submitter — who's proposing it. Default to the current git user (git config user.name). If the user wants to override, accept their input as-is.

One-line description — what would it do? Who's it for? If empty, re-prompt: Give me one sentence about what this idea does. Loop until non-empty.

Don't go deeper than that — this is a lightweight capture, not a spec.

3. Check for duplicates

Before computing an ID or appending anything, fuzzy-match the title against existing backlog entries:

# Normalise a title: lowercase, strip punctuation, collapse whitespace
normalise() { echo "$1" | tr '[:upper:]' '[:lower:]' | tr -d '[:punct:]' | tr -s ' '; }

# Extract existing titles from the backlog table
grep -E '^\| IDEA-[0-9]+ \|' projects/ideas-backlog.md 2>/dev/null \
  | awk -F'|' '{print $3}' \
  | sed 's/^ *//;s/ *$//'

Compare the normalised new title against every normalised existing title using a simple token-overlap heuristic: if ≥ 80% of the words in the shorter title appear in the longer one, flag as a potential duplicate.

If a potential match is found:

⚠ Similar idea already in the backlog:
  IDEA-025 — {existing title} (status: {status})

Is this a duplicate? (y = skip, n = add anyway)
>

• y → skip the append, return without logging anything new, suggest /write-spec IDEA-025 if the user wants to work on the existing idea
• n → continue to step 4

If no match is found, continue silently to step 4.

4. Compute the next ID

# Find the highest existing IDEA-NNN in the backlog file
grep -oE 'IDEA-[0-9]+' <backlog-file> 2>/dev/null | sort -V | tail -1
# Increment by 1, or start at IDEA-001 if none exist

5. Append the entry

If the backlog file doesn't exist, create it with this header:

# Ideas Backlog

Lightweight capture of product ideas, feature concepts, and internal tool proposals.
Use `/idea` to add a new entry. Triage moves entries into `/write-spec`, then into a GitHub Issue.

| ID | Title | Category | Submitter | Date | Status | Description |
|----|-------|----------|-----------|------|--------|-------------|

Append a new row:

| IDEA-NNN | {title} | {category} | {submitter} | YYYY-MM-DD | NEW | {one-line description} |

6. Offer the tracking issue

After the entry is appended, ask:

Would you like me to create a tracking GitHub Issue for IDEA-NNN? (y/n)

If the user says no, skip this step entirely — the backlog entry is already saved, and that's enough.

If yes, resolve the idea body template via the portfolio helper so adopter overrides win when present:

source "$(git rev-parse --show-toplevel)/.claude/hooks/_lib-read-config.sh"
source "$(git rev-parse --show-toplevel)/.claude/hooks/_lib-portfolio-paths.sh"
template=$(portfolio_resolve_template tickets/idea.md)   # → custom-templates/tickets/idea.md if present, else templates/tickets/idea.md

Single-fork adopters (no portfolio block) and adopters with no override fall straight through to templates/tickets/idea.md. Adopters who want a customised idea-tracking-issue shape drop their version at <private_repo>/custom-templates/tickets/idea.md. See templates/README.md for the path-mirroring convention.

Backward-compat fallback: if portfolio_resolve_template returns empty (template file missing — partial adopter setup), fall back to the inline heredoc body below and print a one-line WARN on stderr (WARN: tickets/idea.md template missing — using inline fallback).

Substitute the gathered values into the resolved template (or the fallback heredoc below), then create with the enhancement and idea labels (creating the labels if needed):

gh issue create \
  --title "[Idea] {title}" \
  --body "$(cat <<'EOF'
## Idea
{one-line description}

## Category
{category}

## Submitter
{submitter}

## Backlog Entry
IDEA-NNN — see backlog file.

## Next Step
Triage. Decide whether to spec, schedule, or close.

## Glossary
| Term | Definition |
|------|------------|
| {term} | {definition} |
EOF
)" \
  --label "idea,needs-triage"

Error handling — if gh issue create fails for any reason (missing auth, labels don't exist, network error, rate limit), catch the error and fall back gracefully:

⚠ Couldn't create the tracking issue: {reason}
  The idea is still saved in projects/ideas-backlog.md as IDEA-NNN.

  Try again? (y = retry, n = skip, gh = show the gh error for debugging)
>

Common failure modes and what to do:

Failure	Action
could not resolve repository	Ask the user which repo to file the issue in; the backlog entry was already saved
missing scope: issues:write	Tell the user to run gh auth refresh -s issues and offer to retry
label "idea" not found	Create the label first (gh label create idea) then retry
repository has disabled issues (#653)	GitHub Issues is off on the repo. Surface the fix: gh repo edit <owner/repo> --enable-issues (needs admin), or set tracker.kind to your real tracker. Reuse tracker_check_issues from _lib-tracker.sh to print the hint. The backlog entry is already saved.
HTTP 403 rate-limited	Offer to retry after a short wait
Any other error	Show the raw gh output, skip the tracking issue, keep the backlog entry

The guiding principle: the backlog entry is the primary artefact; the tracking issue is a bonus. Never lose the backlog entry because the GitHub Issue creation failed.

If the issue is created successfully, append the issue URL to the backlog row's Description column as (GH#NN).

7. Offer validation (optional, default-no)

After the GitHub Issue step (whether the user accepted or skipped it), ask:

Validate now? Run /validate-idea IDEA-NNN — y/n (default n)

Default-no respects the lightweight-capture intent of /idea. Most users batch-validate later. If the user accepts, hand off to /validate-idea IDEA-NNN; if they skip, proceed to step 8.

Output

Captured: IDEA-NNN — {title}
Backlog: {file path}
Status: NEW
Tracking issue: {url or "skipped"}
Validation: {"completed — verdict <GREEN|YELLOW|RED>" | "skipped (run /validate-idea IDEA-NNN later)"}

Next: triage with the team, then `/write-spec` if it survives.

Rules

1. Lightweight only — /idea captures, it does not spec. Don't ask for goals, metrics, or requirements here.
2. Always assign an ID — IDEA-NNN, zero-padded to 3 digits.
3. One row per idea — never edit existing rows from this skill; new ideas always append.
4. Status starts at NEW — triage changes it later.
5. Single backlog — every idea goes into projects/ideas-backlog.md at the root of the ops repo; triage assigns it to a project later.
6. Validate before accepting — category must be 1-4; description must be non-empty. Loop until valid; never silently accept garbage.
7. Dedup before appending — fuzzy-match the title against existing entries; flag and confirm before creating a second entry for the same idea.
8. The backlog is the primary artefact — if the tracking issue fails to create, the backlog entry still stands. Never lose data because GitHub was flaky.
9. Don't create the issue silently — always ask first.
10. Never delete — superseded ideas get status SUPERSEDED, not removal.

Status values

Status	Meaning
NEW	Just captured, not triaged
TRIAGED	Reviewed, awaiting decision
SPECCED	/write-spec produced a PRD
SHIPPED	Built and released
WONTDO	Triaged out — not pursuing
SUPERSEDED	Replaced by a different idea

---

Part of ApexYard — multi-project SDLC framework for Claude Code · MIT.