agentleFS
DestrayonDestrayon/Connapse/.claude/skills/next-task/SKILL.md

next-task

Skill22 starsChanged 6 days agoMITView on GitHub

Recommend the best next task to work on based on open GitHub issues. Analyzes priority, dependencies, milestone urgency, and codebase readiness. Trigger when user asks: what should I work on, what is next, next task, pick a task, what to do next, suggest work, prioritize tasks.

10 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 4

    Distinctive

    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. Distinctive:

    Run these in parallel to understand the current state:

    line 12
  2. Very specific:

    Read .claude/state/decisions.md and .claude/state/issues.md for any known blockers, dependencies, or recent decisions that affect …

    line 31
  3. Distinctive:

    If recent merged PRs touched the same area, there's context advantage — the codebase patterns are fresh.

    line 71
  4. Distinctive:

    After presenting, ask the user which one they'd like to start — or if they want to see more options.

    line 111
  5. Very specific:

    Create the branch: git checkout -b <type>/<issue-number>-<short-description>

    line 117
  6. Very specific:

    Read the issue details: gh issue view <number>

    line 125
  7. Distinctive:

    Read the relevant source files mentioned in the issue's implementation notes

    line 126

“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “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

Next Task

Analyze open GitHub issues and recommend the best task to start working on right now.

Step 1: Gather Context

Run these in parallel to understand the current state:

Open issues and board state

gh issue list --repo Destrayon/Connapse --state open --json number,title,labels,milestone,assignees,createdAt,updatedAt --limit 100

Current branch and working state

git status --short
git branch --show-current

Recent activity (what was just worked on)

gh pr list --repo Destrayon/Connapse --state merged --json number,title,mergedAt,labels --limit 5

Check decisions.md and issues.md for blockers or context

Read .claude/state/decisions.md and .claude/state/issues.md for any known blockers, dependencies, or recent decisions that affect prioritization.

Step 2: Score and Rank

Evaluate each open issue against these criteria, in order of importance:

Priority Weight (highest impact)

PriorityScore
P0-Critical100
P1-High70
P2-Medium40
P3-Low10
No priority set20

Milestone Urgency

Issues in the nearest milestone score higher. The reasoning: shipping a complete milestone is more valuable than scattering work across multiple milestones.

MilestoneScore
Current (nearest due date or lowest version)+30
Next+15
Future+5
No milestone+0

Dependency Readiness

An issue that has no blockers and no unmerged dependencies should score higher than one waiting on other work. Check:

  • Does the issue body reference other issues it depends on? Are those closed?
  • Is the issue marked blocked?
  • Would this issue unblock other issues? (check if other issues reference this one)
StatusScore
Unblocks other issues+25
No dependencies, ready to start+15
Dependencies met (referenced issues are closed)+10
Has open dependencies-30
Labeled blocked-50

Momentum Bonus

If recent merged PRs touched the same area, there's context advantage — the codebase patterns are fresh.

ConditionScore
Same area label as a recently merged PR+10
Builds directly on a recently merged PR+15

Size Preference

Smaller issues are easier to start and finish in a single session. Bias toward completing something rather than starting something large.

SizeScore
XS+15
S+10
M+5
L (should be decomposed)-10
No size+0

Step 3: Present Recommendation

Present the top 3 ranked issues in this format:

## Recommended Next Task

### #1: [Issue Title] (#number)
**Score**: [total] | **Priority**: [P-level] | **Milestone**: [version] | **Size**: [size]
**Why this one**: [1-2 sentences explaining why this is the best pick right now —
reference momentum, unblocking, milestone progress, etc.]
**Quick start**: [Which files to look at first, what interface to implement, etc.]

---

### #2: [Issue Title] (#number)
...

### #3: [Issue Title] (#number)
...

After presenting, ask the user which one they'd like to start — or if they want to see more options.

Step 4: Set Up for Work

Once the user picks a task:

  1. Create the branch: git checkout -b <type>/<issue-number>-<short-description>
  2. Move the issue to "In Progress" on the project board:
    # Get the project item ID for this issue
    ITEM_ID=$(gh project item-list 3 --owner Destrayon --format json | jq -r '.items[] | select(.content.number == <ISSUE_NUMBER> and .content.repository == "Destrayon/Connapse") | .id')
    # Move to In Progress
    gh project item-edit --project-id PVT_kwHOAldLE84BQszG --id "$ITEM_ID" --field-id PVTSSF_lAHOAldLE84BQszGzg-vn-U --single-select-option-id 47fc9ee4
  3. Read the issue details: gh issue view <number>
  4. Read the relevant source files mentioned in the issue's implementation notes
  5. Summarize what needs to be done and confirm the approach before writing code

Edge Cases

  • No open issues: Tell the user the board is clear and suggest running /create-tickets if they have ideas to discuss.
  • All issues are blocked: Flag this explicitly and suggest investigating what's blocking progress.
  • Working tree is dirty: Warn before creating a new branch — suggest committing or stashing first.
  • Already on a feature branch: Note that there may be in-progress work. Ask if they want to continue that or switch to something new.

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.