agentleFS
soulmachinesoulmachine/skills/to-features/SKILL.md

to-features

Skill3 starsChanged yesterdayMITView on GitHub

Generate FEATURES.md at the repo root by reading CONTEXT.md and docs/adr/, then enumerating the user-facing features the domain implies. Use after /grill-with-docs has settled the domain language and before /to-prd writes per-feature specs. Bridges the product→engineering gap between domain understanding and feature specification — the missing step that mattpocock's chain doesn't cover natively.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 2 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    After /grill-with-docs has produced or refined CONTEXT.md.

    line 12
  2. Very specific:

    Before /to-prd runs on any individual feature.

    line 13
  3. Very specific:

    If FEATURES.md already exists, treat this as a refinement — propose additions or strikethroughs based on the latest CONTEXT.md and ADRs.

    line 14
  4. Very specific:

    Start with the actor: user-can-…, admin-can-…, guest-can-….

    line 22
  5. Distinctive:

    Don't include implementation details — those land in PRDs downstream.

    line 24
  6. Distinctive:

    Should any features be split or merged?

    line 27
  7. Very specific:

    When a feature ships, mark its line in FEATURES.md with strikethrough + a shipped reference (issue number, PR, or tracker ID).

    line 51

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

to-features

Generate FEATURES.md at the repo root by reading CONTEXT.md and the ADRs under docs/adr/, then enumerating the user-facing features the domain implies. Pairs with /grill-with-docs (upstream, produces CONTEXT.md/ADRs) and /to-prd (downstream, takes one feature → PRD).

When to invoke

  • After /grill-with-docs has produced or refined CONTEXT.md.
  • Before /to-prd runs on any individual feature.
  • If FEATURES.md already exists, treat this as a refinement — propose additions or strikethroughs based on the latest CONTEXT.md and ADRs. Never overwrite shipped (strikethrough) features.

Process

  1. Read CONTEXT.md for domain vocabulary, actors, and invariants.
  2. Read docs/adr/*.md for architectural commitments that constrain what's buildable.
  3. Enumerate user-facing features the domain implies:
    • One feature per line, kebab-case slug + one-line description.
    • Start with the actor: user-can-…, admin-can-…, guest-can-….
    • User-facing, not internal — "User can reset password" not "Add bcrypt hashing."
    • Don't include implementation details — those land in PRDs downstream.
  4. Surface the list to the user for review:
    • Anything missing?
    • Should any features be split or merged?
    • Are slugs unambiguous?
  5. Write FEATURES.md at the repo root once approved.
  6. (Optional, for local-markdown teams) mkdir -p .scratch/<feature-slug>/issues per pending feature to stub the layout downstream stages will fill.

File format

# Features

<!-- One feature per line. Strike through completed features; don't delete. -->

- [ ] user-can-sign-up — A user can create an account with email + password
- [ ] user-can-log-in — A user can sign in to their account
- [x] ~~user-can-reset-password~~ — ~~A user can reset a forgotten password~~ (shipped: #42)
- [ ] user-can-post-photo — A user can upload and share a photo

## Out of scope (rejected)

- ~~user-can-stream-video~~ — Rejected for v1 (see `.out-of-scope/video-streaming.md`)

Discipline: when you finish a feature, strike it through (don't delete)

When a feature ships, mark its line in FEATURES.md with strikethrough + a shipped reference (issue number, PR, or tracker ID). Never delete the line. Strikethrough preserves:

  • Institutional memory — what HAS shipped, not just what's pending.
  • Traceability — link from FEATURES.md to the issue/PR that completed it.
  • Drift resistance — you can't quietly drop a feature; dropping requires explicitly marking it out of scope (and writing the reason to .out-of-scope/).

Transformation on ship:

- [ ] user-can-reset-password — A user can reset a forgotten password
                              ↓
- [x] ~~user-can-reset-password~~ — ~~A user can reset a forgotten password~~ (shipped: #42)

Why this skill exists

mattpocock's chain (/to-prd, /to-issues, /triage) is engineering-side — it assumes features come from product thinking that lives outside the toolchain. swe-workflow needs an explicit bridge between domain understanding (the output of /grill-with-docs) and per-feature specification (the input of /to-prd). That bridge is feature enumeration.

This skill is engineering-adjacent: it reads engineering artifacts (CONTEXT.md, ADRs), but its output (FEATURES.md) is in user-facing language. It's the deliberate seam where product reasoning enters the chain.

When to skip

If your team handles feature enumeration in an external tool — Linear projects, Notion, Productboard, a shared spreadsheet, anything — skip /to-features and use the external source of truth. Pass features directly to /to-prd. This skill exists to fill the seam for teams without an existing product backlog convention, or for solo developers who'd rather keep feature tracking in the repo.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.