AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Next Task Selection

skill-choral-io-choral-skills-next-task-selection · by choral-io

Use when the user asks which accepted Kanban task to work on next and dependencies or readiness may affect the answer.

No reviews yet
0 installs
31 views
0.0% view→install

Install

$ agentstack add skill-choral-io-choral-skills-next-task-selection

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-choral-io-choral-skills-next-task-selection)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
3mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Next Task Selection? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Next Task Selection

Runtime Context

Before acting, resolve ` using the runtime bootstrap rules, then read /.workflow/runtime.md and /.workflow/manifest.yml`; do not assume non-default workflow paths or default ids.

Use this skill to recommend the next accepted delivery task from /planning/KANBAN.md. This skill is read-only by default.

Workflow

  1. Resolve the current member id using /.workflow/runtime.md.
  2. Read Responsibilities and Focus Areas from /members/.md when present; read Availability only when the user asks for capacity-aware selection.
  3. Read local workspace instructions only when the user asks for automatic start, personal execution preferences, or a member-personal recommendation. Use it only as a preference signal.
  4. Read /planning/KANBAN.md.
  5. Read /.workflow/rules/delivery.md.
  6. Read /.workflow/schemas/tasks.md.
  7. Prefer Ready cards over Backlog cards.
  8. Open each candidate's linked task item.
  9. Exclude localized files, local-only notes, archived notes, cards in Blocked, and tasks with unresolved blocked_by entries.
  10. Partition eligible candidates by assignment:
  • tasks where assignees includes the current member id
  • tasks with missing or empty assignees
  • tasks assigned only to other members
  1. Build a lightweight dependency view from blocked_by and related_to; derive downstream unlock potential by reverse-looking up tasks blocked by each candidate.
  2. Score candidates with the selection table in references/scoring.md.
  3. Recommend one next task from the first non-empty partition.
  4. Report blockers, missing metadata, any Ready card that still has unresolved blockers, and any task that looks ready but is not in Ready.

Normalize member and group wikilinks in owners, assignees, and reviewers before matching. For example, [[Gavroche]] and [[Gavroche|Display Name]] both match id Gavroche.

Selection Rules

  • Select only accepted Kanban cards from /planning/KANBAN.md.
  • Do not recommend loose task items from /tasks/*.md that are not linked from a Kanban card.
  • If the user asks to rank task items, backlog candidates, or work not yet on the Kanban board, route to delivery-planning instead of selecting it for implementation.
  • Do not start implementation unless the user explicitly asks.
  • Do not move cards; use kanban-maintenance after maintainer approval.
  • Treat blocked_by as a hard blocker unless all referenced tasks are Done or the blocker is documented as resolved.
  • Treat a Ready card with unresolved blocked_by or readiness: blocked as a board/task consistency problem, not as an eligible candidate.
  • Treat group assignees as team-pool assignment rather than assignment to the current member.
  • Treat owners as durable responsibility, not current assignment.
  • Do not let member profile sections or local workspace instructions override task metadata, dependencies, readiness, approval, safety, or review rules.
  • For source stability, assignment partitions, downstream value, and status-column details, read references/selection-rules.md.

Auto-Start Rules

  • If the user explicitly says the Agent may automatically start, the Agent may start the recommended task from assigned-to-current-member or unassigned according to the user's stated mode.
  • Tasks in assigned-to-others always require a second explicit user confirmation before starting, even when the user allowed automatic start.
  • Automatic start is allowed only when the selected task has readiness: ready, no unresolved blocked_by entries, observable acceptance criteria, committed and remote-synced source material when a default remote exists, and no conflict with current dirty worktree changes.
  • Automatic start must stop if the task appears to require secrets, private customer data, or private personal information.
  • If automatic start is allowed and the selected card must move to Doing, propose or apply the move only under the approval rules in kanban-maintenance.
  • If the user's automatic-start permission is ambiguous, recommend the task and ask before starting implementation.

Output

  • Recommended next task with source link.
  • Short rationale.
  • Candidate score table.
  • Assignment partition used: assigned-to-current-member, unassigned, or assigned-to-others.
  • Auto-start decision: started, needs confirmation, or recommendation only.
  • Blocked or metadata-problem tasks.
  • Suggested board move, if any, as a proposal only.
  • Source stability status: verified, needs verification, or not ready.

References

  • For scoring and metadata examples, read references/scoring.md.
  • For detailed selection, source stability, and auto-start rules, read references/selection-rules.md.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.