agentleFS
rshadershade/finfocus/.agents/skills/promote-issue/SKILL.md

promote-issue

Skill11 starsChanged 13 days agoApache-2.0View on GitHub

Triage the finfocus roadmap backlog and promote the next issue(s) to roadmap/current. Use when the user asks "what should I work on next", "what's next on the roadmap", "promote an issue to current", or wants to triage roadmap/next and roadmap/future before starting unplanned work. Ranks open roadmap/next and roadmap/future issues by a documented scoring rubric (effort, unblocks, epic promotion, triggers), respects suppression rules (parked, timebox/watch, unfired trigger-pending), and presents the top candidates for selection. Selection and promotion only; to claim and implement an issue, use pick-issue. Triggers on: "promote issue", "roadmap promote", "promote to current", "what's ready to work", "triage the backlog".

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 3 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

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

    Read CONTEXT.md first. Any candidate that would cross a CONTEXT.md boundary must be flagged, not silently promoted.

    line 29
  2. Very specific:

    If any query returns exactly its --limit of issues, the result may be truncated.

    line 44
  3. Distinctive:

    If everything is suppressed, present the blocked reasons and the escape hatch only.

    line 125
  4. Distinctive:

    When ROADMAP.md and CONTEXT.md both exist, append the escape hatch as the final option:

    line 128
  5. Very specific:

    Escape hatch selected: follow up with a single AskUserQuestion for a one-line reason, then write into ROADMAP.md Immediate Focus:

    line 136
  6. Distinctive:

    No selection (multi-select only): no file or label changes. Note "no decision — re-prompts next run" in the summary.

    line 157

≈ “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

Promote Issue

Interactive selection of the next issue(s) to work, with optional promotion to roadmap/current. This is the standalone promotion workflow extracted from the /roadmap sync promotion gate — it does not rewrite ROADMAP.md wholesale or sync labels repo-wide.

Preconditions

  • gh CLI authenticated for rshade/finfocus.
  • Repo root has ROADMAP.md and CONTEXT.md. If either is missing, skip every file-update step below and limit yourself to label flips — say so in the summary. The "Deliberately empty" escape hatch only writes ROADMAP.md, so it is unavailable in this case: do not offer it.
  • Read CONTEXT.md first. Any candidate that would cross a CONTEXT.md boundary must be flagged, not silently promoted.

Step 1 — Gather candidates

# Phase labels: roadmap/current | roadmap/next | roadmap/future (mutually exclusive)
gh issue list --state open --label roadmap/next --limit 1000 \
  --json number,title,labels,body,milestone
gh issue list --state open --label roadmap/future --limit 1000 \
  --json number,title,labels,body,milestone
# Cross-check what is already in flight; labels feed the composition_required score
gh issue list --state open --label roadmap/current --limit 1000 --json number,title,labels

If any query returns exactly its --limit of issues, the result may be truncated. Rerun it with a higher limit; if it is still capped, report the truncation in the summary and do not present the ranking as covering the whole backlog.

Excluded from candidacy (never present them):

  • Any label in the noise set: nightly-failure, dependencies, agentic-workflows.
  • roadmap/exclude — dominant; if it coexists with a phase label, report the conflict and leave it to the operator.
  • Already roadmap/current (already in flight; listed only for WIP context).
  • More than one phase label (for example both roadmap/next and roadmap/future, so it appears in two queries). Report the conflict once and leave the issue out of ranking until the operator fixes its labels.

Step 2 — Parse roadmap-meta

Issues may carry structured metadata in an HTML comment in the body:

<!-- roadmap-meta
trigger-pending: <YYYY-MM-DD | git-tag | #issue>
unblocks: 38, 39, 40
epic-parent: 36
gates: 53, 263
spike-parent: 524
verdict: go | needs-clarification | kill | parked
-->

Extract text between <!-- roadmap-meta and -->, then read key: value lines. All fields optional; absent block = "no metadata, always eligible". Adoption is low — every rule below must work when the field is absent.

Step 3 — Apply suppression

A candidate is not workable now if any hold:

  • parked label, or verdict: parked — deliberately not being worked.
  • timebox/watch — gated on a third party; it would block the slot, not clear it.
  • trigger-pending: present and unfired. Fires when: a YYYY-MM-DD today or past; a git tag that exists (git tag --list <value> non-empty); an #issue that is closed. Free-form values never fire automatically.
  • spike-parent: naming another open spike — the parent resolves first.
  • A trigger-pending: naming a milestone (v0.1.0, MVP) suppresses ripeness entirely until the milestone lands — scope decisions are not reversible by scoring.

Suppressed candidates are reported as blocked with the reason, never promoted.

Step 4 — Score remaining candidates

SignalScoreDetection
Fills missing composition_required slot+30Label listed in CONTEXT.md composition_required that no open roadmap/current issue has
Epic-promotion-eligible+15epic-parent: references a closed issue
effort/small+10Label present
effort/medium+5Label present
Has dependents+10Another open issue's unblocks: lists this number
Community-leverage when pipeline thin+8community label AND < 2 community issues closed in last 90d
spec-first but no roadmap-meta block-3Soft nudge
Trigger pending, unfired-20Step 3
Trigger fired+25trigger-pending: condition has arrived
spike gating an effort/large item+12Spike ripens a gated large issue
spike gating an effort/medium item+6Same, medium
timebox/watch-25Step 3

Step 5 — Present candidates

Present the top 3–4 ranked candidates via AskUserQuestion.

  • Label: #NN [S|M|L] label-summary — one-line headline. Top-scored item gets a " (Recommended)" suffix.
  • Description: top two positive contributors, any negative contributor, and a one-line "why this slot" rationale.

Selection mode:

  • roadmap/current has fewer open issues than the WIP target (default target: CONTEXT.md target_focus_depth, else 3) → single-select if one slot, multi-select (max = open slots) if the gate bumped capacity.
  • If everything is suppressed, present the blocked reasons and the escape hatch only.

When ROADMAP.md and CONTEXT.md both exist, append the escape hatch as the final option:

  • Label: "Deliberately empty (release in flight)"
  • Description: "Records 'paused' state. Resume promotion on next run."

Step 6 — Resolve and promote

  1. Escape hatch selected: follow up with a single AskUserQuestion for a one-line reason, then write into ROADMAP.md Immediate Focus:

    *Deliberately empty (YYYY-MM-DD): <reason>. Set by /promote-issue.*

    Replace any existing such annotation in-place.

  2. Candidate(s) selected: for each, flip labels and update the roadmap:

    # Remove every other phase label so the issue carries exactly one
    gh issue edit <n> --add-label roadmap/current \
      --remove-label roadmap/next --remove-label roadmap/future

    Move each entry from its current ROADMAP.md section into "Immediate Focus", preserving description and [S|M|L] indicator, ordered by score (highest first). Append provenance: *Promoted by /promote-issue on YYYY-MM-DD — <rationale>*.

  3. No selection (multi-select only): no file or label changes. Note "no decision — re-prompts next run" in the summary.

Summary output

Report: candidates presented with scores, the operator's decision, issues promoted (with new labels), any suppression/block reasons for near-miss candidates, and any CONTEXT.md boundary flags.

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.