agentsclimarketplace

Daily plan

Skill valentil/featureboard-mcp/skills/daily-plan

Plan today's FeatureBoard work and dispatch it across models. Use when the user says "daily plan", "plan my day", "what should we work on today", "dispatch today's tickets", or wants the board's open tickets budgeted, model-labeled, and started at the right tier.From its SKILL.md

Install
npx -y skills add valentil/featureboard-mcp --skill daily-plan

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

3 things to look at

  • 27 days oldThe repository was created 27 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.2 KB, 730 tokens by cl100k_base, as published. Nobody here has run it

Daily Plan

Build the day plan for a FeatureBoard project (model + effort per ticket), apply it, then dispatch sub-agents on every planned ticket at the right model/effort tier.

1. Build and confirm the plan

  • Call daily_plan with apply: false (pass budgetTokens if the user gave a budget) and show the plan: ticket, model, effort, estimate, and the dispatch groups.
  • Pause for a go-ahead if anything looks off; otherwise continue.
  • Call daily_plan again with apply: true to stamp model:/effort: labels onto the tickets.

2. Dispatch

  • Research first (default ON): for effort:high or research:on tickets not opted out with research:off, call prepare_research and dispatch a haiku/sonnet research sub-agent (its suggestedModel) in parallel; the orchestrator saves each returned brief via add_kb_doc as research/<ticket> BEFORE the implementation dispatch, so those packets carry researchBrief + local ragChunks (BM25, zero-token) automatically.
  • For every ticket in dispatch.parallel (haiku/sonnet/opus): set_status In Progress, get_work_packet, and start a sub-agent at that model with the packet as its brief — these can run concurrently. Right after starting each one, call record_dispatch (worker: "sub-agent", model, parallel: true) so get_agent_monitor and the board show it's running.
  • When parallel tickets touch DISJOINT code areas, give each its own isolated git worktree (create_worktree) so agents don't edit the shared repo at once; merge branches back SERIALLY and cleanup_worktree.
  • Work dispatch.sequential tickets (fable) one at a time, inline in the orchestrator, with a review between tickets; record_dispatch (worker: "sub-agent", model, parallel: false) at start, and again with worker: "orchestrator" when you take a ticket back for review.

3. Effort mapping for each sub-agent brief

  • low — minimal exploration, make the obvious change, verify, stop.
  • medium — normal loop with tests.
  • high — read adjacent code first, consider invariants and back-compat, add tests, self-review the diff before finishing.

4. Orchestrator owns the board

Only the orchestrator writes to the board. As each sub-agent finishes:

  1. Verify its work (run the tests, read the diff).
  2. set_status Done with a one-line completionSummary.
  3. log_work with tokens/additions/deletions and the model used.
  4. commit_feature for that ticket when git is configured.

5. Guardrails and close-out

  • Respect cap:<tokens> labels — wrap up and requeue any ticket about to exceed its cap.
  • commit_feature starts background static checks automatically (pure CPU, no tokens) — keep dispatching, then collect them with get_check_results between tickets and before ending the session; a failed run is a fix-now or file-a-bug before closing out.
  • When the plan is exhausted or the budget is spent, post a day summary to the scratchpad (append_scratchpad) and report to the user.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 326,834. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.