agentleFS
jordanhubbardjordanhubbard/nanolang/skills/roadmap-execution/SKILL.md

roadmap-execution

Skill628 starsChanged 4 months agoApache-2.0View on GitHub

Plan and execute repository product work through docs/ROADMAP.md. Use this before starting features, refactors, optimization programs, or other work with multiple deliverables. Add ordered checkboxes first, execute from top to bottom, and mark only verified work complete.

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    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:

    Read docs/ROADMAP.md and locate the active queue.

    line 17
  2. Very specific:

    Add the requested outcome, implementation work, tests, documentation, and acceptance criteria as ordered - [ ] items before changing code.

    line 18
  3. Distinctive:

    Put dependencies before dependents.

    line 20
  4. Distinctive:

    Create or update GitHub issues for ownership and execution when appropriate.

    line 23
  5. Very specific:

    Discovered bugs always go on the roadmap: If I find a defect, an overclaim ([x] for work that is only a C API, a REPL special-case, a …

    line 29
  6. Distinctive:

    Do not skip a difficult item to implement a more visible dependent feature.

    line 40
  7. Distinctive:

    Add newly discovered required work at the correct dependency position, not automatically at the bottom.

    line 41
  8. Distinctive:

    Keep checkboxes concrete and verifiable.

    line 43
  9. Very specific:

    Keep docs/ROADMAP.md in my first-person voice.

    line 46
  10. Very specific:

    Before ending the session, update roadmap state, update the corresponding GitHub issue, and follow skills/session-completion/SKILL.md.

    line 54

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 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

Roadmap Execution

docs/ROADMAP.md is my product execution order. GitHub Issues is my task ledger. They serve different purposes and I keep both consistent.

Before Implementation

  1. Read docs/ROADMAP.md and locate the active queue.
  2. Add the requested outcome, implementation work, tests, documentation, and acceptance criteria as ordered - [ ] items before changing code.
  3. Put dependencies before dependents. Put measurement before optimization and semantic foundations before surface features.
  4. Mark an existing item - [x] only when repository evidence verifies it.
  5. Create or update GitHub issues for ownership and execution when appropriate.

Small fixes that restore already documented behavior may use an existing roadmap item. If no item describes the work, add one first. Do not turn every typo into a new phase; use the smallest accurate checkbox.

Discovered bugs always go on the roadmap. If I find a defect, an overclaim ([x] for work that is only a C API, a REPL special-case, a layout mismatch), or required work during a session, I add a concrete checkbox in docs/ROADMAP.md at the correct dependency position before I continue. I do not leave defects in chat-only notes. A defect I already fixed in the same session still gets an [x] item so it stays in product history. I file or update a GitHub issue for ownership. Chat is not the ledger.

During Implementation

  • Execute the first unchecked active item whose dependencies are complete.
  • Do not skip a difficult item to implement a more visible dependent feature.
  • Add newly discovered required work at the correct dependency position, not automatically at the bottom.
  • Keep checkboxes concrete and verifiable. Avoid entries such as "improve performance" without workloads and acceptance evidence.
  • Distinguish measured facts, tested behavior, formal proof, and assumptions.
  • Keep docs/ROADMAP.md in my first-person voice.

Completing Items

An item becomes - [x] only after its required code, tests, quality gates, and documentation are complete. A partial implementation stays unchecked or is split into smaller honest items.

Before ending the session, update roadmap state, update the corresponding GitHub issue, and follow skills/session-completion/SKILL.md.

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.

Reviews can't be read right now.

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.