agentleFS
zajcazajca/pohunek/.agents/skills/plan-phase/SKILL.md

plan-phase

Skill8 starsChanged 9 days agoMITView on GitHub

Plan the next pohunek phase interactively — review the roadmap and current state, resolve open questions ONE at a time with the user deciding, then record a complete end-to-end plan as a GitHub issue (no PoC, no shortcuts). Use when the user asks to plan or design the next phase or milestone before implementing it.

5 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

1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

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

    Ground yourself in the current state: Read before proposing: docs/ROADMAP.md, docs/phases/, docs/architecture.md (authoritative scope), the …

    line 31

“Common” is measured: kept by at least 6 in 10 of the 4 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

plan-phase — plan the next phase into a GitHub issue

Produces the spec that the milestone skill later implements: a complete, end-to-end plan recorded as a GitHub issue, arrived at through an interactive question-and-decision pass with the user. Planning is where this project refuses shortcuts — the point is a full design, not a minimal proof of concept. The plan is persisted in the issue via the github-workflow skill; no local NEXT.md is written.

Steps

  1. Resolve the target issue. Route through the github-workflow skill's issue-resolution rules: use the issue URL/number the user or task gives; otherwise deduplicate against existing issues first. When the user hands over concrete new phase scope and no matching issue exists, auto-create the planning issue and add it to the configured project without a further ask; ask only when the scope is genuinely ambiguous or competing issues both plausibly cover it. Create the issue when planning starts (so decisions land somewhere live as they are made) and grow the body as questions are settled.
  2. Ground yourself in the current state. Read before proposing: docs/ROADMAP.md, docs/phases/, docs/architecture.md (authoritative scope), the relevant docs/design/*.md (accepted technical constraints stay authoritative until a deliberate decision revises them — then the issue must say so), and the current open issues that touch the same surfaces. Skim the crates the phase will touch so the plan is grounded in the real code, not assumptions.
  3. Frame the phase. State what this phase is for and where it sits in the roadmap. List the key assumptions explicitly. Respect the repo's hard constraints (owner-first direct operation, owner browser path, additive optional relay, PTY/TUI-first, remote over NetBird not SSH, providers shell-out and client-only, no back-compat shims).
  4. Resolve open questions — one at a time. This is the core of the skill and how the user works: for each unresolved question, ask it on its own, describe what the problem is, and offer 2-3 concrete options with trade-offs. Wait for the user's decision before moving to the next question. Do not batch every question at once, and do not silently pick an answer. Record each decision and its rationale in the issue body as it is settled.
  5. Write the complete plan into the issue. The issue body (via the github-workflow skill) covers the phase end to end: scope, the design and decisions recorded above, the crates/surfaces affected (protocol ripples into client/daemon/cli and the generated sdk/ts/ types if the wire changes; UI work belongs in zajca/pohunek-work, never in this repository), knowledge-bundle and docs/public-api.md impact, and an explicit, testable definition-of-done list with stable IDs (D1, D2, ...) the milestone and milestone-review skills will check against, and a proposed PR stack: ordered slices, one coherent concern each, mapped to DoD items (per pullRequests in .github/agent-workflow.json). Ensure the issue is in the configured project for delivery tracking. Being tracked as Todo does not imply the proposed design is accepted; acceptance is an explicit decision recorded on the issue.
  6. Confirm. Summarize the plan and the DoD, link the issue, and note that it is ready to implement (via the milestone skill).

Hard rules

  • No PoC, no minimal versions, no shortcuts unless the user explicitly asks for a reduced scope. The plan must describe the full solution.
  • Best solution over fastest — never trade correctness or completeness for implementation speed during planning.
  • The plan does not authorize implementation or merge. It defines what was agreed; building it still goes through the milestone skill, and landing it through merge-advance/pr-handoff.
  • Planning goes here as a GitHub issue. Design docs that outlive a milestone belong under docs/design/ only when the user asks to store them there; do not create local NEXT.md or RFC files as work authorities.

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.