skills/track-roadmap/SKILL.md
Plan, audit, and resume project roadmaps. Use when asked to "create a roadmap", "generate a roadmap", "what should we build next", "brainstorm features", or "audit the roadmap".
npx skillsauth add antjanus/skillbox track-roadmapInstall 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.
Maintain ROADMAP.md in the project root as a living, high-level feature plan, so decisions about what to build next are intentional, not reactive. Keep it to features (user-visible capabilities), not tasks — use track-session for task-level work. A healthy roadmap is 5-15 committed features readable in under 2 minutes.
| Mode | Command | Essence |
|------|---------|---------|
| Generate | /track-roadmap or generate | Offer a codebase scan, ask 4 discovery questions, propose a feature list, then write ROADMAP.md after the user confirms |
| Update | /track-roadmap update | Read current file, ask what changed (add/remove/complete/reword), apply after confirmation |
| Audit | /track-roadmap audit | Scan code to mark each feature Done/In-Progress/Not-Started/Unclear, review relevance with the user, update |
| Brainstorm | /track-roadmap brainstorm | Divergent ideation — explore directions before committing; viable ideas land in "Future Ideas" as status:idea |
| Resume | /track-roadmap resume | Check session state, present remaining features, user picks one, hand off to /track-session |
Full per-mode procedures (discovery questions, brainstorm question banks, audit steps): reference/MODES.md.
---
schema: cc-dash/roadmap@1
project: project-name-here
description: One-line project purpose
last_updated: YYYY-MM-DDTHH:MM:SS-TZ
---
# Roadmap
> Project purpose in one sentence.
## Core Features
<!-- category:core -->
- <!-- id:r_XXXXX status:planned --> **Feature Name** - What it does and why it matters.
- <!-- id:r_XXXXX status:in-progress started:YYYY-MM-DD --> **Feature Name** - Short description.
## Future Ideas
<!-- category:future -->
- <!-- id:r_XXXXX status:idea --> **Feature Name** - Short description.
## Completed
<!-- category:completed -->
- <!-- id:r_XXXXX status:done completed:YYYY-MM-DD --> ~~**Feature Name**~~ - Short description. *(Completed: YYYY-MM-DD)*
Format rules (the cc-dash schema is parsed by the dashboard — don't omit the markers):
schema, project, description, last_updated.id (r_ + 5 random [a-z0-9], permanent — never change) and a status (planned / in-progress / done / idea).<!-- category:slug --> on the next line. Categories are flexible.✅ Good — frontmatter, unique IDs, statuses, category slugs, strikethrough + date on completed items:
## Core Features
<!-- category:core -->
- <!-- id:r_k8x2m status:done completed:2026-01-15 --> ~~**Task CRUD**~~ - Create/read/update/delete tasks with title, description, due date. *(Completed: 2026-01-15)*
- <!-- id:r_m3p7q status:in-progress started:2026-02-01 --> **Task lists** - Organize tasks into named lists (Work, Personal, Shopping).
❌ Bad — - tasks / - lists / - make it look good / - fix bugs: no descriptions, no groupings, mixes features with tasks, no IDs or purpose.
Full ✅/❌ comparisons for generate, audit, update, and resume modes: reference/EXAMPLES.md.
track-session to plan and track the implementation; SESSION_PROGRESS.md references the ROADMAP item ID.r_xxxxx roadmap issue.generate → pick a feature → /track-session → build
brainstorm → explore ideas → update → commit to plan
resume → check session → pick feature → /track-session → build
audit → review progress → update → adjust plan
Troubleshooting (too many features, over-granular features, decision paralysis, resume edge cases): reference/TROUBLESHOOTING.md.
development
EXPERIMENTAL. Mine recent Claude Code transcripts for friction events, cluster them by active skill, propose patches for skills with 3+ friction events, validate each patch via headless replay, scrub the report through /publish-check, and present an EVOLUTION_REPORT.md for human review on a branch (never auto-merge). Use when asked to "evolve my skills", "audit skills against recent friction", "propose skill improvements from transcripts", "run the skill evolution pipeline", or as part of a weekly skill-quality cadence.
testing
Manual QA tracking — things tests can't verify. Use when asked to "create a QA list", "set up QA for this project", "what should I QA", "track manual QA", "audit the QA list", or "start manual QA".
development
Multi-source web research with cited synthesis in chat. Use when asked to "research X", "deep research on Y", "deep dive on Z", "investigate this topic", "compare X and Y", "pros and cons of X", or "survey the landscape of Y".
development
Use this skill whenever the user wants a multi-agent review of local changes — triggers include "review my code", "review these changes", "do a code review", or "check my changes before I commit". Writes REVIEW.md. Do NOT use for an open PR by number (use /review) or a security-specific pass (use /security-review).