agentleFS
sssemilsssemil/SKILLS/skills/brutal/brutal-plan/SKILL.md

brutal-plan

Skill2 starsChanged 4 months agoMITView on GitHub

Orchestrate repo-grounded grilling, specification review, cohesive ticket design, and work-store publication. Use when planning a feature, refactor, migration, or system change into a decision-complete parent plan and executable tasks through BRUTAL.md integrations.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    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 ../brutal-shared/integration-resolver.md, resolve the work store, and load its support module.

    line 14
  2. Very specific:

    Read ../brutal-grill/SKILL.md, ../brutal-spec/SKILL.md, and ../brutal-tickets/SKILL.md.

    line 16
  3. Very specific:

    Read repository rules from AGENTS.md, CLAUDE.md, TARGET.md, and their referenced workflow documents.

    line 18
  4. Distinctive:

    If a completed investigation map is supplied, read its decisions and assets and retain a link to it in the specification.

    line 20
  5. Distinctive:

    Use the normal route when material requirements and design decisions can be resolved in the current planning conversation.

    line 25
  6. Distinctive:

    Use these markers for normal plan and task dedupe:

    line 48
  7. Distinctive:

    Before each write, search by source, parent, and exact title.

    line 55

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Brutal Plan

Drive planning through explicit stage gates, then publish the approved plan and tasks through the work store configured in BRUTAL.md. Keep the analysis rigorous and the user-facing checkpoints compact.

Required Context

  1. Read ../brutal-shared/integration-resolver.md, resolve the work store, and load its support module.
  2. Read ../brutal-grill/SKILL.md, ../brutal-spec/SKILL.md, and ../brutal-tickets/SKILL.md.
  3. Read repository rules from AGENTS.md, CLAUDE.md, TARGET.md, and their referenced workflow documents.
  4. If a completed investigation map is supplied, read its decisions and assets and retain a link to it in the specification.

Choose The Route

Use the normal route when material requirements and design decisions can be resolved in the current planning conversation. Read ../brutal-wayfinder/SKILL.md and offer that route when a necessary decision depends on separate research, a prototype, unavailable access, or more than one planning session. Explain the reason and confirm the route before persisting an investigation map.

Run The Gates

  1. Apply brutal-grill until its requirements brief is decision-complete.
  2. Apply brutal-spec, resolve every validated PLAN BLOCKER, and present a compact gate-one summary: outcome, scope, key decisions, material risks, and acceptance criteria. Obtain explicit approval. Show the full specification only when the user asks for it.
  3. Apply brutal-tickets and present a compact gate-two graph with one line per cohesive scope: delivered outcome and genuine blockers. Obtain explicit approval of scope boundaries, granularity, blocking edges, and coverage. Show complete ticket bodies only when the user asks for them.
  4. Publish only after both gates pass. A rejection or unresolved decision leaves all work-store artifacts and domain documents unchanged.

Publish Resumably

Use these markers for normal plan and task dedupe:

Source: brutal-plan
Legacy sources: brutal-plan, linear-brutal-plan, gitlear-brutal-plan

Before each write, search by source, parent, and exact title. Create or update the parent and approved glossary or ADR changes before exposing implementation tasks. Stage complete child bodies in blocker-first order, wire relationships, update the parent with final identifiers, then promote the children into the intake queue. If publication fails, stop, keep unpromoted work outside intake, report every completed and missing write, and resume from that inventory.

The parent body contains the approved specification, completed-map reference if any, child graph, plan-level acceptance criteria, and known risks. Every child body follows the brutal-tickets contract and mirrors its parent and blockers even when native relationships exist.

Publish through the resolved work-store contract. Create/update the plan parent before exposing children, use the adapter's native staging state or its prepare-before-create strategy, create complete children blocker-first, wire native relationships where supported, and mirror every parent/blocker in the body. A remote adapter never creates local work-store artifacts; local resumable state is allowed only where its support module specifies it.

Final Response

Report the work store and parent artifact, list each task in dependency order on one line, name the material risk or partial failure if one exists, and identify the next unblocked task. Link to persisted detail instead of repeating it.

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.