agentleFS
anthony-chaudharyanthony-chaudhary/fak/.agents/skills/milestone-score/SKILL.md

milestone-score

Skill49 starsChanged 14 days agoApache-2.0View on GitHub

One repeatable pass that makes the milestone report's CLIMB and ROADMAP retirable by the RSI loop — the milestone counterpart of quality-score (code) and stability-score (trust under iteration). Runs `fak milestone-scorecard --json` over the report's OWN two dimensions — the M0..M7 support-maturity climb shortfall PLUS the un-progressed tracked-epic roadmap gaps — folds them into one deterministic milestone_debt integer + a worst-first milestone_worklist, and retires debt worst-first. Use when this named workflow matches the task.

The file

Files
SKILL.md

/milestone-score — make the milestone climb + roadmap retirable, then retire it

What this does. The milestone report (fak milestone report) RECORDS the project's position — the M0..M7 maturity grid and the tracked-epic roadmap — but a trend report does not grade a retire-able debt, so the RSI loop could not drive it like it drives code-quality or stability. This pass gives the report its scorecard surface: a deterministic milestone_debt integer and a worst-first worklist, then retires that debt one real climb or epic-close at a time.

It is an instance of the generic scorecard doctrine pointed at one surface: the milestone program's climb to MATURED and its roadmap convergence. The tool is the Go verb fak milestone-scorecard (cmd/fak/milestonescorecard.go over internal/milestonereport), and it folds into tools/scorecard_control_pane.py next to the rest of the family — registered as milestone (debt milestone_debt) alongside the milestone_climb ratchet gate.

The two dimensions (and why this is NOT a parallel maturity scale)

milestone_debt = climb_debt + roadmap_debt, folded from the SAME two dimensions the report already folds — no re-derived scale:

  • climb — climb_debt = sum over every cell below the MATURED floor (M4 Correct) of (M4 - current_rung) rung-steps. One defect per owed rung-step, so the kernel's len(Defects) count fold equals the rung-weighted shortfall. This is the report's OWN matured-floor framing (raw distance-to-MATURED across the grid), NOT the declared-target/regime-ceiling framing support_maturity_debt uses — so the two scorecards stay distinct: support-maturity fences each cell to its regime ceiling and scores the declared-target shortfall; milestone scores the headline project climb to a single MATURED bar. They COMPOSE, they do not duplicate.
  • roadmap — roadmap_debt = sum over the MEASURED DISCRETE tracked epics of their open children (total - closed). An ongoing PROGRAM (kernel-optimization, cache-optimization) has no 100% so it never owes roadmap debt; an UNREADABLE epic is excluded — you cannot retire what you cannot measure, and that gap is the report's ACTION verdict's job, not the scorecard's.

The worst-first milestone_worklist orders climb cells by rung shortfall (lowest rung, biggest gap, first) ahead of roadmap gaps by open-child count — the climb is the deeper, structural debt, so it retires first.

Run it as an RSI pass (the five steps)

  1. Run it — fak milestone-scorecard --json for the control-pane payload (read corpus.milestone_debt + corpus.milestone_worklist); fak milestone-scorecard for the rendered snapshot. Save a baseline first: fak milestone-scorecard --json > /tmp/milestone-before.json. The roadmap half resolves each tracked epic's children live via gh; to fold hermetically (offline / no gh), pass --epics-from <file> with a data file carrying a pre-resolved counts block.
  2. Retire milestone_debt worst-first — drain corpus.milestone_worklist top-down: climb the lowest-rung cell up to MATURED (M4) — fence it, run it, then prove it correct — then close the most-open DISCRETE epic's children. Never move a cell to M4 by re-labeling; a rung climb is witnessed by the cell actually meeting the rung's bar.
  3. Re-pin the climb ratchet on a REAL climb — the two witnessed climb KPIs (matured_cells, milestone_progress) are pinned in docs/milestones/baseline.json and a regression in EITHER reds CI via the milestone_climb control-pane gate. After a genuine climb improvement, re-pin: fak milestone-scorecard --pin. Never --pin to silence a regression — that ratchets the floor DOWN.
  4. Re-measure + prove — fak milestone-scorecard --compare /tmp/milestone-before.json prints the debt delta; the worklist must have shrunk and milestone_debt must not have risen. There is no per-surface markdown snapshot to regenerate (the report itself is the snapshot); the proof is the --compare delta plus the re-pinned baseline.
  5. Commit only the milestone lane, by explicit path — run fak sync reconcile --apply, then commit via fak commit --path:
    fak sync reconcile --apply
    fak commit --path docs/milestones/baseline.json --path <the cell/epic evidence you moved> -m "<subject>" [--push]
    fak sync push
    End the subject with the (fak milestonereport) trailer (or (fak <leaf>) for the concrete code a rung climb touched). Never git add -A.

The anti-gaming law

Retire a defect by moving the real cell or closing the real epic, never by editing the grid labels or the tracked-epic set to flatter the number. A climb defect is retired by a cell genuinely meeting the next rung's bar (witnessed by the report's own collection), and a roadmap defect is retired by a child issue actually closing — not by un-tracking the epic or redefining MATURED. If "fixing" a defect would mean relaxing a rung's definition or dropping an epic from the tracked set, stop — that is not a real retirement.

Re-pinning the climb baseline (handle with care)

The climb ratchet (milestone_climb, debt climb_ratchet_debt) is a DISTINCT gate from milestone_debt: a cell count that holds while milestone_debt rises still reds the control pane here. Do not blindly --pin after a milestone-score pass: confirm matured_cells did not decrease and milestone_progress did not regress FIRST, then re-pin only the genuine climb. A blind re-pin blesses a regression the way a portfolio re-pin hides a per-metric one.

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.